Wikipedia:Village pump (technical)/Archive 93

Wikipedia talk:Miscellany for deletion#Overlap

Would an editor be able to help at Wikipedia talk:Miscellany for deletion#Overlap (permanent link)? Thanks, Cunard (talk) 04:04, 5 September 2011 (UTC)

  Resolved
 – Bug fixed. mc10 (t/c) 00:58, 6 September 2011 (UTC)

I found a bug while examining Wikipedia:Help desk#Generated confirm email link appears to be broken (404 - File Not Found). The secure server currently sends out email confirmation links starting with http://en.wikipedia.org/wikipedia/en/wiki/Special:ConfirmEmail. This looks like a hodgepodge of a normal and secure link, and it gives a 404 error. It should be either http://en.wikipedia.org/wiki/Special:ConfirmEmail or https://secure.wikimedia.org/wikipedia/en/wiki/Special:ConfirmEmail. Can somebody fix it or report it in a better place? Thanks. PrimeHunter (talk) 15:18, 30 August 2011 (UTC)

I have submitted it as bugzilla:30647. PrimeHunter (talk) 00:59, 31 August 2011 (UTC)
The bug has been fixed. PrimeHunter (talk) 03:38, 1 September 2011 (UTC)

usernames

I've been attempting to create an account here but have been thwarted by the lack of any clear guidelines on usernames. Existing users seem to have a wide variety of creative names, but my ASCII-only attempts result in "You have not specified a valid user name", without any indication of what is wrong. 111.233.10.80 (talk) 10:56, 2 September 2011 (UTC)

The software allows the full Unicode character set to be used, so that editors from Russia or Japan, say, can choose names in their own language. However, a few characters are not allowed, because they would cause problems when pages are displayed. According to this page, the following characters are not allowed: # < > [ ] | { } and _. Are you being mislead by the decorative signatures that some editors use? For example, there's a user here named "Orangemike" who signs as <font color="darkorange">Orange Mike</font>. The font tag is part of his signature, not part of his username. -- John of Reading (talk) 12:14, 2 September 2011 (UTC)
(edit conflict) I'm not sure - perhaps the username is already in use? If you have difficulty you can try the Wikipedia:Request an account process instead. — Martin (MSGJ · talk) 12:16, 2 September 2011 (UTC)
I don't see where in Wikipedia:Naming conventions (technical restrictions)#Forbidden characters it specifies that underscores are disallowed, nor where it says that any of these guidelines apply to usernames. And I have no idea how I was expected to find this information. I imagine other potential users would find it helpful if this were incorporated into the signup page in some way. Nonetheless thank you for letting me know what the problem was. = ) 111.233.10.80 (talk) 12:42, 2 September 2011 (UTC)
It was removed as part of the outreach:Account Creation Improvement Project. Apparently people hate reading instructions which is fine if they'd improve the error messages. — Dispenser 12:57, 2 September 2011 (UTC)
Spaces, underlines and non-breaking spaces are treated equivalently in page names (black hole, black_hole and black hole being the same article); that ought to apply to user names too (or underlines forbidden altogether) otherwise in principle two users could find themselves sharing a user page. I'll try to log in using underlines instead of spaces in my user name, now...
A. di M.plédréachtaí 19:26, 4 September 2011 (UTC)
It works!
A. di M.plédréachtaí 19:32, 4 September 2011 (UTC)

Just for the notice: two underscores/whitespaces (so Black__Hole) are technically forbidden... mabdul 05:45, 6 September 2011 (UTC)

Own subpage created too quickly: how can it be discarded / deleted

Hello to the administrator or user who would like to help a lost ± newbie!

I am sorry if here is the wrong place to ask this, but I didn't find a Wikipedia page that is exactly adequate to deal with this problem:
I thought I was on Wiktionary, and I was surprised not to find the Russian word язык (language). I was ready then to create the article, as the message proposed it. Then I clicked too swiftly on a very practical link (when needed), but it was too late when I realized that 'Язык' had a capital, which may not be done on Wiktionary (except when you create an article about a proper name or an acronym). This 'user subpage' has the following url: http://en.wikipedia.org/w/index.php?title=User:Air_Miss/%D0%AF%D0%B7%D1%8B%D0%BA&action=edit&preload=Template:Article_wizard/userpageskeleton&editintro=Wikipedia:Article_wizard/Wizard-New_edit_instructions_userdraft

Could any wizard help me, simply deleting this user subpage I don't need? I will send this wizard a friendly kiss through the air for helping! --Air Miss Ѡ 09:46, 5 September 2011 (UTC)

User:Air Miss/Язык does not exist; it hasn't been saved. - David Biddulph (talk) 09:53, 5 September 2011 (UTC)
Right. Perhaps you were confused by seeing prefilled content in the edit box for your url http://en.wikipedia.org/w/index.php?title=User:Air_Miss/%D0%AF%D0%B7%D1%8B%D0%BA&action=edit&preload=Template:Article_wizard/userpageskeleton&editintro=Wikipedia:Article_wizard/Wizard-New_edit_instructions_userdraft. This content is loaded from Template:Article_wizard/userpageskeleton by preload=Template:Article_wizard/userpageskeleton in the url. You didn't click Save page so nothing was saved. PrimeHunter (talk) 11:55, 5 September 2011 (UTC)

But should you need it in the future, try Wikipedia:Speedy#User pages - X201 (talk) 12:07, 5 September 2011 (UTC)

Huggle error

My Huggle script has noticed me twice times today, that “an error occurred and needs to close”. What could possibly cause this error? Alex discussion 17:27, 5 September 2011 (UTC)

Typically this page is used for discussion of bugs in Wikipedia. I think you should report this at WT:HG, which might get more Huggle developers to see it. Thanks, Nathan2055talk - review 18:05, 5 September 2011 (UTC)

Template:User Wikipedian for

I just noticed on my user page the text in this template misses a space. It reads "This user has been on Wikipedia for 9 yearsand 12 days." I have no idea how to fix this. Besides, the template appears to be full-protected. Anyone know what's wrong with it?--Atlan (talk) 23:06, 5 September 2011 (UTC)

I messed it up. It's now been fixed (I hope). Ucucha (talk) 23:33, 5 September 2011 (UTC)
That did the trick, thanks.--Atlan (talk) 23:36, 5 September 2011 (UTC)

copy of Template:ARSHa

Can someone with admin powers temporarily undelete this page for me? I just need to see the coding. thank you.

http://en.wikipedia.org/w/index.php?title=Template:ARSHa&action=edit&redlink=1

Okip 22:14, 3 September 2011 (UTC)

Sure. :) It's at User:Okip/ARSHA. When you're finished with it, you can just retag it. --Moonriddengirl (talk) 17:01, 6 September 2011 (UTC)

Gadget not working

I've noticed yesterday and today that the "Improved diff view" gadget (listed under Preferences, Gadgets, Editing, and described as part of wikEd) is not working for me. I use Firefox 6.0.1. When I click on the green icon below the default diff display, nothing happens. Any help? --Tryptofish (talk) 19:48, 5 September 2011 (UTC)

It has already been reported at User talk:Cacycle/wikEd#Error message about local diff script. Further discussion belongs there. PrimeHunter (talk) 22:12, 5 September 2011 (UTC)
As reported there, the issue has now been fixed. PrimeHunter (talk) 22:52, 5 September 2011 (UTC)
Thanks. --Tryptofish (talk) 17:10, 6 September 2011 (UTC)

Abuse filter doesn't include ebay listings?

Call me crazy but I can't see any reason not to automatically reject edits like this. Equazcion (talk) 01:04, 6 Sep 2011 (UTC)

The spam blacklist can prevent such edits. MER-C 10:41, 6 September 2011 (UTC)

Substitution problem

I created a template, User:UcuchaBot/FAS line, that is intended to be used substituted on WP:FAS; all templates and magic words called should also be substituted. However, substituting currently doesn't work, in that #expr throws an error in the fourth cell. See this example:


|- |Aug 2011 | 3730 | 6627 | Expression error: Unrecognised punctuation character "{"% | 32 | 4 | style="background: #FFE3E3; color: black; vertical-align: middle; text-align: center; " class="table-no2" |28 | 43 | 105

Does anyone know whether there is a way of using subst: or perhaps safesubst: that fixes this? Ucucha (talk) 01:20, 6 September 2011 (UTC)

The problem may be that {{subst:FA number}} does not produce a number but a formatnum which evaluates to a number. PrimeHunter (talk) 06:09, 6 September 2011 (UTC)
Yes, I think you need another (includeonly'ed) "subst:" (or maybe more) for inside {{FA number}}.--Kotniski (talk) 11:11, 6 September 2011 (UTC)
Or no. When deleting the subst: from that cell the error cell #7, the problem stays. I don't know the calling template. The cell is constructing a template with params. I'd try:
  • Check value of incoming params FAs promoted, FAs demoted. (is there a value at all, and is it numeric?) If logic allows, write {{{FAs promoted|0}}}-{{{FAs demoted|0}}}.
  • Check whether incoming templateused is OK (a template name)
  • Replace the template pipe with {{!}}
-DePiep (talk) 11:24, 6 September 2011 (UTC)
You seem to be looking at the seventh cell. I was talking about the fourth cell.--Kotniski (talk) 11:32, 6 September 2011 (UTC)
You are right. I moved my eyes because #4 looks OK now, and now #7 gives the #expr:-error. Which problem are we to solve now? -DePiep (talk)
Seems #4 looks OK on the template page itself, but not when the template is substituted - hopefully the addition of subst:'s inside {{FA number}} will fix it. As to #7, it generates an error on the template page, but possibly might work on substitution, provided the right parameters are fed in (though I'm somewhat sceptical...) --Kotniski (talk) 11:42, 6 September 2011 (UTC)
(ec) #7 gives an error in the template version only, which I don't care about. #4 is the important one, since it gives an error when the template is substituted (its intended use). I think PrimeHunter and Kotniski are right about the underlying problem, so I'll add some subst:-ing capabilities to FA number. Thanks all for the help. Ucucha (talk) 11:43, 6 September 2011 (UTC)
Right. I'm off for a cup of tea then. -DePiep (talk) 12:01, 6 September 2011 (UTC)

Edit box citation feature

  Resolved

How it will appear to you will depend on the skin you're using and possibly any user-scripts you've added, but in File:EditBox.jpg the top row has a "Cite" toggle. As you can see, it's been activated, so the line below includes a "templates" drop-down. If you select that, you get a list of templates, and selecting one of those gives you a dialogue box to complete.

What is this set of features called, and where do we go, to propose changes to it? Andy Mabbett (Pigsonthewing); Andy's talk; Andy's edits 09:51, 6 September 2011 (UTC)

This is WP:RefToolbar. ---— Gadget850 (Ed) talk 09:59, 6 September 2011 (UTC)
So it is. Thank you. Andy Mabbett (Pigsonthewing); Andy's talk; Andy's edits 11:07, 6 September 2011 (UTC)

Broken thumbnail

Hi! I've uploaded a picture of my own work at http://en.wikipedia.org/wiki/File:Cyclopentolate_1%25_Pupils.jpg#file, however the thumbnail appears to be broken. If you click on the broken thumbnail the image loads correctly. Not sure if it's something I've done wrong (most likely!) Any ideas would be greatly appreciated. Thanks! Ilovebaddies (talk) 17:51, 6 September 2011 (UTC)

Weird. Maybe the percentage symbol in the name is causing problems? I don't feel like there'd be anything in the image itself that is causing problems. Gary King (talk · scripts) 21:03, 6 September 2011 (UTC)
When I go the thumbnail's URL (http://upload.wikimedia.org/wikipedia/en/thumb/0/07/Cyclopentolate_1%25_Pupils.jpg/800px-Cyclopentolate_1%25_Pupils.jpg), I get the following error: "Error generating thumbnail Error creating thumbnail: convert: unable to open image `/mnt/thumbs/wikipedia/en/thumb/0/07/Cyclopentolate_1%%_Pupils.jpg/800px-Cyclopentolate_1%%_Pupils.jpg': @ error/blob.c/OpenBlob/2498." Notice the double % in that path; I think the percentage sign may indeed be the culprit. Ucucha (talk) 21:05, 6 September 2011 (UTC)
Indeed, that'll do it. I've moved the file and spelled out "percent", and that took care of it. Avicennasis @ 21:38, 7 Elul 5771 / 21:38, 6 September 2011 (UTC)

bugzilla:30789. Ucucha (talk) 22:37, 6 September 2011 (UTC)

How did italics show up in the article title?

The article is New York City Serenade (film). But I was the only contributor until a bot came along to mark it uncategorized. I'm not sure how to fix that, but I didn't put the italics in the title.Vchimpanzee · talk · contributions · 21:00, 6 September 2011 (UTC)

Those are produced by {{Infobox film}} using {{Italic title}}. Ucucha (talk) 21:02, 6 September 2011 (UTC)
Okay, thanks.Vchimpanzee · talk · contributions · 21:20, 6 September 2011 (UTC)

Non-existent image appears at Mayawati

Can anyone explain how does this non-existent image appear at the Mayawati article? Much obliged. Dr.K. λogosπraxis 00:02, 7 September 2011 (UTC)

The link was in the infobox ... I've removed it.Tagishsimon (talk) 00:09, 7 September 2011 (UTC) --
I know where the filename was. But how can a non-existent filename render in an infobox? Am I missing something here? In other words, what server is this phantom image located in and how was it retrieved by the infobox? Dr.K. λogosπraxis 00:18, 7 September 2011 (UTC)
No image appeared when I viewed the article. Are you saying that an image rendered, or a filename rendered? If the former, maybe something in your browser's cache? If the latter, that's what happens when you specify a filename fora non-existent file. --Tagishsimon (talk) 00:22, 7 September 2011 (UTC)
It was an actual image. Not a filename. Removing redlinked filenames from infoboxes is easy. Mayawati appeared in a political poster. I don't recall seeing this image before but maybe you are right and it was my browser's cache. I'll clear it and then try to see if it renders in the old revision, before you removed the filename. Thank you. Dr.K. λogosπraxis 00:28, 7 September 2011 (UTC)
The file was recently (within a few hours) deleted from commons, see [1] ΔT The only constant 00:32, 7 September 2011 (UTC)
Thank you Δ. It now appears as a redlink. It no longer renders as an actual pic, even though I did not clear my browser's cache. Dr.K. λogosπraxis 00:36, 7 September 2011 (UTC)

Skin

Can somebody tell me on monobook how to change the background colour of the skin from grey white streaky to dark blue and the side text white? I just want the frame of the page to be much darker, the top strip and the side panel. When I change the background colour in my browser it paints the whole page that colour. I want to retain the white page but with a dark background to make it stand out Does somebody have any coding to do this?♦ Dr. Blofeld 19:53, 30 August 2011 (UTC)

You can edit your monobook.css page, and add CSS rules there. I'm not sure what you mean by "side text white", but to change your background color, something like the following should work:
body {
    background: #00008B; /* Or a similar color */
}
Hope that works. mc10 (t/c) 01:04, 6 September 2011 (UTC)

I mean if you have a dark border you will need white text so you can read it in contrast..♦ Dr. Blofeld 11:14, 7 September 2011 (UTC)

How to add extra buttons to the toolbar (for common use templates)

There are some templates I use very frequently (welcome, for example). I'd like to be able to add them with one click. Is there any script/feature that would allow me to tie buttons (tabs, whatsnot) into templates I'd define? I asked for this feature to be added to wikEd but since I got no reply there for weeks, I need to search elsewhere. --Piotr Konieczny aka Prokonsul Piotrus| talk to me 17:23, 4 September 2011 (UTC)

Yes, see my monobook for 'extraeditbuttons'. –xenotalk 17:45, 4 September 2011 (UTC)
User:MarkS/extraeditbuttons.js? I thought those died years ago, I was using them before they were obsolete by the some MediaWiki updates. I even suggested it here a while ago that somebody should take it over, but nobody did. Hmm, let's see if it works now... --Piotr Konieczny aka Prokonsul Piotrus| talk to me 20:33, 4 September 2011 (UTC)
Sadly, the extraeditbuttons seem as dead as I remember. I added them to my vector skin, but I still see the unchanged edit toolbar. I tried disabling the editing toolbar in the preferences, but that did nothing. And Xeno, looking at your monobook, I see you've imported the script but are not doing anything with it? I think one needs (needed) to play with those settings to see changes. --Piotr Konieczny aka Prokonsul Piotrus| talk to me 20:43, 4 September 2011 (UTC)
I currently have two custom buttons, one that says *{{notdone}} ~~~~ and another that says *{{done}} ~~~~. –xenotalk 03:50, 5 September 2011 (UTC)
I tried adding your version of the script to mine ([2]) but no effect. I wonder why? Could it be that it doesn't work under vector? Has anybody got it working on vector? Is it conflicting with some other script or setting? --Piotr Konieczny aka Prokonsul Piotrus| talk to me 17:40, 7 September 2011 (UTC)
Did you refresh your cache? I don't know if it will work in vector. –xenotalk 17:48, 7 September 2011 (UTC)

Request for protection

Hi! I didn't know where to ask this, but could someone put protection from IP for this article - Kārlis Skrastiņš - and maybe other Lokomotiv players too? Source. They maybe aren't death so it could be good to make a protection. Thanks.--Edgars2007 (Talk/Contributions) 13:49, 7 September 2011 (UTC)

I semi-protected it. In the future you should ask for protection on WP:RFPP. Ruslik_Zero 14:38, 7 September 2011 (UTC)
OK, thanhs!--Edgars2007 (Talk/Contributions) 14:40, 7 September 2011 (UTC)

Statistics progrma not updated

I'm not sure why but it seems like the Article statistics program isn't displaying the updated data from Sept 1 forward. The developer of the app (Henrik) Hasn't made an edit since March 2011 so I thought I would leave it here instead. Here is a link to the tool with an example. --Kumioko (talk) 18:46, 6 September 2011 (UTC)

Weirdly, http://stats.grok.se/en/201109/United_States does not show past the 1st, but http://stats.grok.se/en/latest/United_States shows right up through today's. Pi.1415926535 (talk) 21:47, 6 September 2011 (UTC)
Yeah, I noticed also the stats have stopped for a while. Henrik may not have contributed as a user since March, but he or somebody is reading his email. Usually, if somebody emails Henrik, he gets it fixed. --Maile66 (talk) 23:04, 6 September 2011 (UTC)
Ok thanks Ill send him something today. --Kumioko (talk) 14:48, 7 September 2011 (UTC)
It seems to be updating again, but I see there are some discrepancies between United States (201109) and United States (latest): the "latest" figures are all one day ahead of the "201109" figures. —Bruce1eetalk 11:13, 8 September 2011 (UTC)

Disabling rollback button on watchlist

I've had numerous instances where I accidentally clicked the rollback button on my watchlist. It's particularly obnoxious when I sign in to check my watchlist on my smartphone. Is there currently a tool to disable it? Magog the Ogre (talk) 21:41, 7 September 2011 (UTC)

.page-Special_Watchlist .mw-rollback-link {display:none} in Special:MyPage/vector.css will do the trick. EVula // talk // // 21:48, 7 September 2011 (UTC)
special:mypage/skin.css for non-vectorites. –xenotalk 22:15, 7 September 2011 (UTC)
Ah, I couldn't remember the theme-independent name. Thanks. EVula // talk // // 16:26, 8 September 2011 (UTC)

Is there a way to trigger that in JavaScript? I'm thinking of only disabling it on my phone, and I'll need a JavaScript call to ascertain my browser version. Magog the Ogre (talk) 23:34, 7 September 2011 (UTC)

Possibly you'll be able to do that with CSS alone, because MediaWiki associates some classes with the html element based on the client. Something like
.client-phone .page-Special_Watchlist .mw-rollback-link {display:none}
, where ".client-phone" is the class that your phone gets, and which you should be able to find in the HTML source. Ucucha (talk) 01:53, 8 September 2011 (UTC)

Finding Empty categories

Is there an easy want to take a parent category and check for empty sub-categories recursively? Avicennasis @ 23:32, 8 Elul 5771 / 23:32, 7 September 2011 (UTC)N

Maybe you are looking for something simpler, but if you click on all the little + signs next to the subcategory names, it will show the number of pages and categories within each next level subcategory. --After Midnight 0001 01:25, 8 September 2011 (UTC)
I'm looking for something more automated.   For something like Category:Stub categories, I can't imagine clicking through 10k+ sub-categories that way. :-) Avicennasis @ 01:31, 9 Elul 5771 /
You could ask someone like MZMcBride to do a database report for you. However, I suggest you look at Wikipedia:Database reports/Empty categories, noting the excluded phrases listed at the top: from today's report, I see only one empty stub category (Category:Uruguayan football defender stubs) and one empty stub-related category (Category:Uncategorized stubs from August 2011). — This, that, and the other (talk) 11:17, 8 September 2011 (UTC)
That's nifty, and useful. I've requested this report on Commons.   Now, are there any tools that can sort subcats by number of pages in them? kinda like this, but with less manual labor on my part? Avicennasis @ 12:10, 10 Elul 5771 / 12:10, 9 September 2011 (UTC)

Error 403

I am getting an error 403 on my smartphone and cannot log in--Woogie10w (talk) 11:44, 9 September 2011 (UTC)

OK now signed in on a secure server--Woogie10w (talk) 12:23, 9 September 2011 (UTC)

Ayuda - Help

Si hay algún usuario que sepa idioma español, respóndame. El problema es que en Wikipedia en español tengo un usuario de nombre Ferdinand. Como ocurre con varios usuarios en español, al irse estos a la Wikipedia inglesa, sus usuarios existen. Resulta que no me ocurre eso porque en esta Wikipedia hay un User:Ferdinand que no ha tenido participación y está inactivo. Solicito su borrado. Tengo un usuario creado en la inglesa pero cuando entro con su cuenta, me cambia la cuenta en español y eso es mucha molestia. Muchas gracias :) --190.81.168.163 (talk) 23:42, 9 September 2011 (UTC)

Tenemos una página, WP:CHUU, para solicitudes como la suya. He pedido que un bureaucrat cambie el nombre del usuario Ferdinand de en.wikipedia, para que usted puede usar la cuenta de Ferdinand en en.wikipedia. Creo que usted tendría que ir a es:Especial:Fusionar_cuenta_global y crear un cuenta SUL. Ucucha (talk) 02:44, 10 September 2011 (UTC)

User Contributions missing from Toolbox

In the left-hand column under Toolbox, I haven't been able to get User Contributions to display, even when I remove Expand Citations or any add'l gadget added there. The problem is in both IE9 and Firefox. Is there a bug out there? --Funandtrvl (talk) 01:29, 10 September 2011 (UTC)

Just wanted to add that it seems that User contributions is now under the tabs on the top right. It maybe due to a gadget that I have checked in "My preferences", but at least, it's found now. --Funandtrvl (talk) 03:23, 25 September 2011 (UTC)

Idea for new cite template.

I've noticed that there is no template to cite a film or movie. Could such a template be added? This might be particularly helpful for documentaries. NewManOfAnOldAge (talk) 01:31, 9 September 2011 (UTC)

{{Cite video}}, compliant with Citation Style 1. ---— Gadget850 (Ed) talk 02:36, 9 September 2011 (UTC)
Why doesn't this show up under the Cite template list that is accessible whenever you are editing a wikipedia page? I see on that page a ton of cite templates but under this template list you can only cite web, news, book and journal. NewManOfAnOldAge (talk) 20:48, 9 September 2011 (UTC)
You can discuss that at WP:Reftoolbar. ProveIt GT supports most of the CS1 templates; enable it by setting Preferences → Gadgets → ProveIt. I added tool support to Citation Style 1. ---— Gadget850 (Ed) talk 06:55, 10 September 2011 (UTC)

Installation of extension

Following on from Wikipedia:Village pump (technical)/Archive 73#Obtaining image dimensions, Wikipedia:Village pump (technical)/Archive 78#Image height?, also other threads but most recently Template talk:Multiple image#image height, how would we go about getting mw:Extension:MediaFunctions installed on English Wikipedia? --Redrose64 (talk) 23:54, 10 September 2011 (UTC)

You need: a) community consensus that we want it; and b) a sysadmin willing to turn the extension on (I don't know how robust the code is). The way to find such a sysadmin is by filing a bugzilla report. Ucucha (talk) 00:04, 11 September 2011 (UTC)
Where would the appropriate place be to start a discussion to try for consensus?--Taylornate (talk) 01:22, 11 September 2011 (UTC)
WP:Village pump (proposals), I suppose. Ucucha (talk) 01:40, 11 September 2011 (UTC)

Template for regular tasks reminder

Is there a template for something like a regular tasks reminder? E.g. [[User:Me/ToDo]]

{{Reminder
|Page     =Category:Some category
|ToDo     =Clean this or that
|Interval =Sundays
}}

If I visited [[User:Me/ToDo]] on a Monday there would a link to Category:Some category reading "Clean this or that". The date stuff is no problem with the help of magic words and parser functions. The problem is marking the task as done so it disappears until next Sunday.

The extension could be expanded to use "monthly", "Xth of the month", etc. Anyone got an idea how to properly handle it? --Subfader (talk) 16:51, 10 September 2011 (UTC)

Yes, using something like {{#ifeq: {{CURRENTDAYNAME}} | Sunday | [[:Category:Some category|Clean this or that]]}}, or some other magic word at mw:Help:Magic words#Date and time. Ucucha (talk) 20:55, 10 September 2011 (UTC)
As I wrote that's not the problem. The problem is that it should appear until the task is done. If I come back online on Monday and not Sunday, the task should wait for me. If done, it should disappear. Next Sunday a new task appears. --Subfader (talk) 11:44, 11 September 2011 (UTC)
What would be the switch between done/to-be-done? Category is empty? Editor (you) edits the templated page every week? -DePiep (talk) 12:01, 11 September 2011 (UTC)
#switch is just an extended #ifeq and would mean the template would need to read all possible date configrations already. My example was simplified. Actually it'd also useful for reminding me on things like checking external sources without RSS regulary, for stuff like new images etc. --Subfader (talk) 13:09, 11 September 2011 (UTC)
I think DePiep was asking how the template could tell whether a task is done or to be done. I can't see an easy way to do that. Ucucha (talk) 13:35, 11 September 2011 (UTC)
Correct, Ucucha. -DePiep (talk) 13:39, 11 September 2011 (UTC)
I see. Makes sense now :) Ok, I give up. Thanks for the replies tho! --Subfader (talk) 21:04, 11 September 2011 (UTC)

Diff formatting

Viewing the source code that produces the diff:

http://en.wikipedia.org/w/index.php?oldid=449815434&diff=prev

One will see

<span class="diffchange diffchange-inline"><ref name="T28 on RETROBrick"/> </span>

Can someone talk about when this new '"class" was introduced, and when it appears?

It used to be the case that

<span class=\"diffchange\">some text here</span>

Was used to format all red text (and still is for the most part). Lately, though, I've been seeing the "diffchange diffchange-inline" format pop-up every once in a while -- and this was causing some issue with my scripts. Thanks, West.andrew.g (talk) 13:03, 11 September 2011 (UTC)

Perhaps with the introduction of MediaWiki 1.17, a few months back. If you use something like $(".diffchange"), your script should work no matter which of the two formats is used. Ucucha (talk) 13:19, 11 September 2011 (UTC)
Yeah, no big deal to work around. This was just the cause of a very deep bug in WP:STiki that caused my parsing of the diff to break. Thanks, West.andrew.g (talk) 13:27, 11 September 2011 (UTC)

No edit buttons for me on one particular page?

I am suddenly unable to see the section edit links at WP:FEED but I have them on every other WP page. And yes, I do have the 'section edit' option ennabled in my preferences. I already cleared my cache, have refreshed the page many times, have closed the page and then opened it in a new window, have tried it in FIrefox and in Safari...still no edit buttons on WP:FEED. Thanks, Shearonink (talk) 03:05, 10 September 2011 (UTC)

The main page is full-protected. You'll need to find the correct subpage. Strange Passerby (talkcont) 03:10, 10 September 2011 (UTC)
I don't believe he was trying to edit the protected page. He was trying to edit the individual subpages via the section edit links that should be on the main WP:FEED page but they were not there for him. -- œ 03:53, 10 September 2011 (UTC)
As far as I know, if a page is protected from edits by a user then there will never be section edit links, not even when the section is transcluded from a page the user is able to edit. It might be confusing if some sections had edit links but not others or the whole page. If you click the "View source" tab then the bottom of the window will show a list of transcluded pages. Click on a page there and then you will have section edit links if you can edit the page. However, the protection of WP:FEED should probably be removed when it causes this problem. PrimeHunter (talk) 04:18, 10 September 2011 (UTC)
Indeed, that's the case - the full-protection will suppress the section edit links from transcluded pages for non-admins. Avicennasis @ 04:34, 11 Elul 5771 / 04:34, 10 September 2011 (UTC)
Well jeez.. that would probably be the reason could be part of the reason (i just noticed it was only recently protected) why WP:FEED is so neglected! if only admins are able to easily provide feedback.. -- œ 06:21, 10 September 2011 (UTC)

Hmmm, I see that you are right, that the edit links aren't there is you go to the main page. The edit links are there is you go to a particular day's page, such as Wikipedia:Requests for feedback/2011 September 10. However, a casual editor interested in providing feedback may not know to go to the day page. I can assure you that the protection is not the main problem, but I see that it could contribute.

However, there are a few problems. I didn't design the setup, someone else did, so I have only a limited understanding of how the page works. I thought it made sense to do it with the transclusions, but I now see a downside. One problem is that removing the protection won't help, as the intention is to edit the underlying page, not the transcluded page. So while I can remove the protection, or anyone else can, that will just encourage people to edit the transcluded page, and that's why it was fully protected. That said, with Protection removed, non admin could click on the section edit, which is what they are supposed to do, and someone will just have to clean up if someone tries to edit the whole page. Maybe one of the more knowledgeable readers can suggest a better option. Ideally, the whole page should be fully protected, but clicking on the edit button for an individual section would trigger the edit of the underlying section, than the transcluded section. However, feel free to change the protection to semi, if you think that is better than the current situation.

Another problem is that I've basically given up on Feedback. I was supplying much of the feedback, either directly, by doing it myself, or indirectly, by begging at the help desk for volunteers. I've written essays, and tried to get the welcoming committee and others interested, and frankly, there's just no interest, so I've given up. Maybe a better approach is to just shut it down.--SPhilbrickT 12:44, 10 September 2011 (UTC)

Well, one could "fake" edit links for the days at least, to get people to the right page. Something like this should do it. Avicennasis @ 13:45, 11 Elul 5771 / 13:45, 10 September 2011 (UTC)
  • I've reduced it to semiprotection for now, just to make sure that people have some way of responding. Obviously, this is an interim measure, but it seems to me a better interim measure than leaving it unusable. :/ (Scary that it took so long for anybody to notice. People really don't give feedback. :() --Moonriddengirl (talk) 16:18, 12 September 2011 (UTC)

Weird obscene BLP image vandalism at Nelson's Collared Lemming

I can't see this in Chrome, IE 9 or FF6 but I've got an OTRS ticket that says that using Chrome on two different IPSs and with several computers there's an obscene image and that "The whole page is a hotlink to "lawlhwut.feenode.net" which chrome blocks as suspect malware." The image, an attack on User:NawlinWiki, does not move when you scroll (says the email). Dougweller (talk) 08:39, 10 September 2011 (UTC)

I don't know how old the ticket is, but one of the called templates on that page was vandalized within the past few weeks. It's revdel'd, so I can't tell if that vandalism is the same one they are talking about. It could be something stuck in the cache - I would have them try opening the page in Google Chrome Incognito (which shouldn't load anything from Chrome's cache.) That's all I can think of. Otherwise, it loads fine on my Chrome as well. Avicennasis @ 10:51, 11 Elul 5771 / 10:51, 10 September 2011 (UTC)
New last night. I don't seem to be able to view it either, so I guess Oversight got involved. But he used several computers - the first time he wasn't at home, then he went home and used several computers there. Very odd. Dougweller (talk) 12:33, 10 September 2011 (UTC)
The vandalism was discussed at Wikipedia:Administrators' noticeboard/IncidentArchive717#Dodo. It did indeed cause what the OTRS ticket says. It sounds like a version using the 21 August version of the template was still cached somewhere. PrimeHunter (talk) 15:13, 10 September 2011 (UTC)
Thanks. Weird though, two ISPs and several computers used during the last 36 hours showed the image. Dougweller (talk) 18:33, 10 September 2011 (UTC)

Another report of this just came up at User talk:Stemonitis#Palaeomonetes antrorum. What I suspect is that the attacker purged as many pages as he could after vandalizing Template:IUCN, so that the template still shows up there (though not for us logged-in users, as we usually don't see cached pages). Therefore, it may be necessary to have all pages using Template:IUCN purged to avoid this thing showing up for anonymous users. Ucucha (talk) 23:15, 11 September 2011 (UTC)

I'll set a bot loose on it to have all the pages updated.   Avicennasis @ 05:29, 13 Elul 5771 / 05:29, 12 September 2011 (UTC)
  Done Hopefully that should take care of the issue. Avicennasis @ 11:11, 13 Elul 5771 / 11:11, 12 September 2011 (UTC)
Thanks, I hope that solves it. Ucucha (talk) 11:23, 12 September 2011 (UTC)

Horizontal scroll wierdness

I sort of remember seeing this before, but if you compare File:SCW reported ref problem.png, which is what I see, with File:Refs_of_Spanish_Civil_War2.jpg (another user), you'll notice that some – possibly importantly the last in their sections – go right off to the right. I'm on firefox, I think the other user (from the look of it) Google Chrome. There's nothing in the obvious wikimarkup to explain this. How can I fix it? Grandiose (me, talk, contribs) 20:50, 10 September 2011 (UTC)

I can confirm that the bug occurs in Chrome 13.0.782.220 as well as Safari 5.0.6 on Mac OS X 10.5. Ucucha (talk) 20:59, 10 September 2011 (UTC)
I have tried rearranging the refs but the last one ALWAYS sticks out in Spanish Civil War#Further Reading. – Plarem (User talk contribs)   21:05, 10 September 2011 (UTC)
It's a webkit bug which happens when a column width if forced wider then the available space. Since 60em is unlikely to break into columns anyway, I've removed those widths. Edokter (talk) — 21:08, 10 September 2011 (UTC)
Thank you, now we can pass the Spanish Civil War to GA. – Plarem (User talk contribs)   21:12, 10 September 2011 (UTC)
Are you aware that WP:SIG#Images prohibits images in signatures? Ucucha (talk) 22:09, 10 September 2011 (UTC)
Didn't we discover this a while back? ---— Gadget850 (Ed) talk 16:24, 12 September 2011 (UTC)
If you mean the problem that started the thread, probably. I couldn't think of a search term to scour the archives with. Grandiose (me, talk, contribs) 18:55, 12 September 2011 (UTC)

How to enable SOCKS proxy in Huggle?

Some wikipedians access wikipedia via SSH, in other words, they must use SOCKS proxy... Can anyone tell me if there is a direct way to enable SOCKS proxy in Huggle?  Thanks! - Dr. Cravix ★Daydream Nation 01:48, 11 September 2011 (UTC)

I just checked Huggles' Project Page and noticed you put a request for this there too. I guess this is   Done here for now. Best to ask there if Huggle's developers can help you with this, methinks. JguyTalkDone 14:28, 12 September 2011 (UTC)

Wikipedia:Miscellany for deletion

Wikipedia:Miscellany for deletion seems to be having some .css issues, didn't know who to tell, thought you guys could fix it or pass it to the right place. Thanx. Mlpearc powwow 02:55, 10 September 2011 (UTC)

What exactly is the problem? It looks fine to me. Ucucha (talk) 02:57, 10 September 2011 (UTC)
Sorry must be my browser. The "Centralized discussion" infobox is overlapping "Deleting pages in other people's userspace" options. Mlpearc powwow 03:36, 10 September 2011 (UTC)
Looked fine to me in Opera, IE, and Chrome, but in Firefox I saw the overlay you were talking about. I've fixed it with this edit.   (At least, it now renders correctly on my Firefox.) Let me know if that helped. Avicennasis @ 04:41, 11 Elul 5771 / 04:41, 10 September 2011 (UTC)
Not keen on that - the {{clear}} has added about six inches of white space for me (Firefox 3.6.15) -- John of Reading (talk) 07:21, 13 September 2011 (UTC)
Err. Yes, what John said. Killiondude (talk) 07:28, 13 September 2011 (UTC)
Ok. I played around and did some editing in Firefox, and I *think* I got it this time. Let me know if the recent version is any better. Avicennasis @ 11:50, 14 Elul 5771 / 11:50, 13 September 2011 (UTC)
Much better. There's no overlap until the window is made unreasonably small. -- John of Reading (talk) 11:58, 13 September 2011 (UTC)

Bits/Jquery

I'm trying to track down a problem which is causing web filtering software to block some Javascript on Wikipedia. The link being blocked is

http://bits.wikimedia.org/en.wikipedia.org/load.php?debug=false&lang=en&modules=jquery%7Cmediawiki&only=scripts&skin=vector&version=20110912T172820Z

What's being blocked appears to be any JS gadgets and JS I have added to my vector.js. I'm not too familiar with the internals of Mediawiki or Javascript etc. and was wondering what else Bits/Jquery provides that may be affected? (It's entirely possible that I haven't noticed other problems yet, as this one was discovered by chance.) Thanks, --Kateshortforbob talk 12:21, 13 September 2011 (UTC)

What is "&cad" used for in google.books.com url?

I've searched Google's documentation on this (and failed to find anything), and removing it from URLs seems make no difference at all.

See

vs

So does it do anything, or is this some junk parameter only of use to Google? (If this is better suited for the REFDESK, just move this thead there). Headbomb {talk / contribs / physics / books} 18:16, 13 September 2011 (UTC)

Google Books seems to have a lot of parameters without any obvious use; I tend to remove all parameters except id and pag (or whatever it is called) when I add links to it. Ucucha (talk) 22:45, 13 September 2011 (UTC)
Using the API has most of the fields, but I don see cad. ---— Gadget850 (Ed) talk 23:01, 13 September 2011 (UTC)
Actually it lacks most of them. The url structure really is poorly documented. Headbomb {talk / contribs / physics / books} 23:52, 13 September 2011 (UTC)

As part of a thread at WP:HD, I just went to Wikipedia:CAPINYOASS, where I got one of the "Please take the 2011 Wikipedia Mobile Readers Survey by clicking here." dropdowns, similar to the messages telling me that we're about to be electing stewards and things like that. Since I'm editing the primary version of Wikipedia, not the mobile version, any ideas why I got this message? I've never used the mobile version, except while using it on a friend's phone, and I was logged out at the time. Nyttend (talk) 03:43, 14 September 2011 (UTC)

I just got the above message on a desktop using FF 3.6 (geesh, I didn't realize how un-updated this was). Killiondude (talk) 04:55, 14 September 2011 (UTC)
I have never edited from a mobile device. I am getting the dropdown over and over (FF 6, mostly looking at article histories). I would love a cookie to stop this - or a real cookie to perk me up if my browser can't have one. 71.234.215.133 (talk) 05:22, 14 September 2011 (UTC)
I poked the person who activated it here. Killiondude (talk) 05:30, 14 September 2011 (UTC)
Hey guys. It's not a survey to be displayed on mobile devices, it's a survey asking readers on non-mobile devices about their usage patterns on a mobile device. The reason it's not displayed on mobile devices is that no one would sit there for 15 minutes on a cell phone taking a survey when they just popped on to mobile wikipedia to look up one tiny thing on the go, and we wouldn't get much meaningful data if we put up a 2-question survey for mobile devices Nimish Gautam (talk) 17:14, 14 September 2011 (UTC)
I actually had wondered that after I posted on your talk page on meta. Thanks for the clarification. Killiondude (talk) 20:37, 14 September 2011 (UTC)

A hyphen is not a hyphen

Recently I was editing and in the edit box the punctuation for a list appeared to be a hyphen, and it looked exactly like what appeared when I used the hyphen on my keyboard to add to that list. But the result looked different. My hyphen was shorter than the other ones after the edit, so the only thing I knew to do was copy and paste the other type of hyphen in the list where I had just edited.Vchimpanzee · talk · contributions · 21:00, 14 September 2011 (UTC)

Dash#Common dashes --Tagishsimon (talk) 21:03, 14 September 2011 (UTC)
It's probably &ndash but – looks identical to the hyphen in the edit box in this case. Well, not here, but there.Vchimpanzee · talk · contributions · 21:16, 14 September 2011 (UTC)
Okay, now that I know, there is a slight difference here but I don't see how those longer hyphens were made to appear in the first place. And here is the diff so you'll know where to look.Vchimpanzee · talk · contributions · 21:19, 14 September 2011 (UTC)
(edit conflict) When you are editing, below the "Save page", "Show preview" and "Show changes" buttons, you should see a drop-down menu. If this does not state either "Insert" or "Wiki markup", select either of those from the menu. Then, to the right of that menu, you will see some symbols. The first two of these are – and —; these are the en-dash and em-dash respectively, and you can click on them to insert them into the edit box at the current cursor position. The character found on keyboards is known as the hyphen-minus; it looks the same as the en-dash in an edit window, but is slightly shorter on the finished page. --Redrose64 (talk) 21:21, 14 September 2011 (UTC)

Show diff size in contribs list?

When I'm viewing my contribution list (or someone else's for what matters), would it be possible to view the diff size in addition to the rest of the stuff via some monobook.js customization? (It could be made a preference setting too.) It's done on watchlists, so why not on this one too? Headbomb {talk / contribs / physics / books} 17:55, 14 September 2011 (UTC)

Interesting question but I have a sad response for you: it can't be done. At least, currently.
Diff sizes are stored temporarily for use in recent changes and watchlists. Your contributions are pulled from a different data source - an older one - and the diff changes aren't stored there. It might be possible to add this, but doing so would be difficult to backport. You can always open a bug about it (please do - I want this feature), but I know for a fact that it is impractical right now (I wanted to have it from Global Profile--Jorm (WMF) (talk) 19:29, 14 September 2011 (UTC)
Done. https://bugzilla.wikimedia.org/show_bug.cgi?id=30903 Headbomb {talk / contribs / physics / books} 19:33, 14 September 2011 (UTC)
ty!--Jorm (WMF) (talk) 20:44, 14 September 2011 (UTC)
  3 users like this. 20:56, 14 September 2011 (UTC)

Woah, apparently this has been fixed. I suppose it only needs to be reviewed and rolled out? Headbomb {talk / contribs / physics / books} 20:56, 15 September 2011 (UTC)

Yes, it needs to be reviewed, and then merged into a WMF branch before it gets rolled out. This will probably take a couple months, so don't get super exited just yet.--Jorm (WMF) (talk) 22:10, 15 September 2011 (UTC)

Infobox musical artist appears to be seriously whacked

It looks like someone changed image size handling in {{infobox musical artist}}, breaking it badly in at least some use cases.

See Dave Wakeling for an example; specified image now displays full size (and a full size it is), rather than as a thumbnail.

If this isn't the right place to report this, please let me know!--NapoliRoma (talk) 13:50, 15 September 2011 (UTC)

Never mind (I think). I see the bug is actually in the use of the template on that page.--NapoliRoma (talk) 14:00, 15 September 2011 (UTC)
For future ref, problems with templates should normally be notified at the template's talk page (in this case Template talk:Infobox musical artist). On this page, you should find the discussion relating to changes to the template, which may turn out to be the best place for your comments. --Redrose64 (talk) 17:03, 15 September 2011 (UTC)
Thanks. In retrospect I can see I probably was over-reacting, but I was looking for a "big red button," because my first impression was that a widely-used infobox template had a serious break, and I wasn't sure if posting to the template page would catch anyone's attention in a timely manner.--NapoliRoma (talk) 19:48, 15 September 2011 (UTC)

Multiple new editors, using same connection

I'm currently in the middle of leading an outreach even, with another starting in a couple of hours. We're in public venue, using their wi-fi, and are hitting limits on accounts created; plus edit filters stopping trainees from completing legitimate edits. Any suggestions, please? Andy Mabbett (Pigsonthewing); Andy's talk; Andy's edits 15:10, 15 September 2011 (UTC)

I've granted your account the accountcreator flag; you will have to be logged in and create the accounts for them by visiting special:CreateAccount. –xenotalk 15:20, 15 September 2011 (UTC)
Saved my life! Thank you. Andy Mabbett (Pigsonthewing); Andy's talk; Andy's edits 15:56, 15 September 2011 (UTC)

Article feedback tool + WikiLove feature

What are the plans regarding the Article Feedback tool and the WikiLove feature. Are these features now officially part of the English Wikipedia? Or are we still in some kind of "trial phase" regarding these features? I am sorry if I missed the discussions regarding these questions, thus if there were discussions regarding this, can someone point me to a place where I can see them? Toshio Yamaguchi (talk) 14:32, 14 September 2011 (UTC)

I do not know if the AFT is permanent on Wikipedia at this point, but I can personally attest that the WMF is still working on successfully utilizing data it generates. One of my projects is reaching out to some of the responders. :) If nobody pops in here with more info, you may be able to get more at the talk page of mw:Article_feedback. So far as I know, WikiLove has been officially part of the English Wikipedia since the day it was rolled out, but you may be able to request clarification of that at mw:WikiLove. --Maggie Dennis (WMF) (talk) 17:48, 16 September 2011 (UTC)

Template:International Democrat Union

The template is not wrapping. It displays full width requiring the horizontal scroll bar. – Lionel (talk) 09:03, 15 September 2011 (UTC)

Template:International Democrat Union (edit | talk | history | links | watch | logs)
It's working OK for me, using Windows 7 and Firefox 4.0.1. What browser and version and you using? -- John of Reading (talk) 15:56, 15 September 2011 (UTC)
IE 8. Fyi this one, which is similar, works fine {{Alliance of European Conservatives and Reformists}}Lionel (talk) 07:08, 16 September 2011 (UTC)
Yuck! Yes, the IDU template looks awful in IE9. Perhaps this is because the IDU template uses {{nowrap}}, while the AECR template uses non-breaking spaces? -- John of Reading (talk) 07:28, 16 September 2011 (UTC)
Well, as if we needed more template drama, I just noticed it doesn't look like the non-breaking spaces are working at AECR. On my browser the flag for Electoral Action is one one line, and the link on the next! – Lionel (talk) 08:12, 16 September 2011 (UTC)

Loging in using smartphone?

I get 403 error. I noted the message above and tried using secure server, but it won't log me in globaly and there's no link to secure server on language version I want. Plus it logged me out here allready, although I selected remember me option. Any advice?89.219.142.38 (talk) 10:13, 16 September 2011 (UTC)

Lost access to my talk page

I can no longer access my talk page User talk:Simon Burchell; I keep getting the following message:

A database query syntax error has occurred. This may indicate a bug in the software. The last attempted database query was:

(SQL query hidden)

from within function "User::deleteNewtalk". Database returned error "1205: Lock wait timeout exceeded; try restarting transaction (10.0.6.46)".

Any help would be appreciated. Regards, Simon Burchell (talk) 14:16, 16 September 2011 (UTC)


I am having the exact same issue. Worse, I can't get rid of the "You have new messages" bar at the top - neither User talk:Orderinchaos nor the direct link from the bar work, though strangely I can load *other* people's talk pages fine. Have tried on Firefox, Chrome and Opera to see if Firefox was the problem. Orderinchaos 14:27, 16 September 2011 (UTC)
I got that message (Simon's one) when I left a note at User talk:Canley (only user talk page I've edited recently). The edit went through fine, but it took ages and then that message came up, making me think it hadn't. This is using Safari. Very odd. Jenks24 (talk) 14:33, 16 September 2011 (UTC)

I've tried both Firefox and IE, no joy with either. Simon Burchell (talk) 14:38, 16 September 2011 (UTC)

I just got the following when trying to edit User talk:Ucucha with my alternative account User:Oryzomys:

A database query syntax error has occurred. This may indicate a bug in the software. The last attempted database query was:

(SQL query hidden)

from within function "User::updateNewtalk". Database returned error "1205: Lock wait timeout exceeded; try restarting transaction (10.0.6.46)".

I'll open a bugzilla ticket about these issues. Ucucha (talk) 14:38, 16 September 2011 (UTC)

bugzilla:30931. Discussion on IRC just now suggests this has been fixed. Ucucha (talk) 14:51, 16 September 2011 (UTC)
(edit conflict) I don't think you'd need a bugzilla, these seem to just be database issues that should be resolved in due course. Could ping someone in #wikimedia-tech. –xenotalk 14:51, 16 September 2011 (UTC)
You were right, Xeno (though someone else had already filed it on bugzilla). Apparently, the issue was that someone was running a MySQL database query on a user talk-related table in the enwiki database that was messed up. Due to something related to subselects and http://bugs.mysql.com/bug.php?id=46947, that locked part of the database. Domas fixed the issue by killing that query. Ucucha (talk) 14:57, 16 September 2011 (UTC)
It's not fixed. I'm still getting:

A database query syntax error has occurred. This may indicate a bug in the software. The last attempted database query was:

(SQL query hidden)

from within function "User::updateNewtalk". Database returned error "1205: Lock wait timeout exceeded; try restarting transaction (10.0.6.46)".

whenever I try to add a block notification to somebody's talk page. However, the edit does get saved...maybe the "You have new messages (last change)" isn't showing up for them? Reaper Eternal (talk) 15:23, 16 September 2011 (UTC)

Can somebody fix wp:FFD?

I does not properly transclude the last few days, might have something to do with the extreme number of nominations. Some tech savy editor on a fast pc who can take a look? Yoenit (talk) 14:43, 16 September 2011 (UTC)

Probably could be solved just by linking, rather than transcluding, the 2011-09-11 subpage. –xenotalk 14:49, 16 September 2011 (UTC)
It is the extreme number of nominations that's causing it. There is a limit on the number of parser functions that can work on a single page. The nom's use 4 functions a piece, I believe, and with so many nominations, it adds up. Once the list shrinks down a bit, though, the problem should resolve itself. Avicennasis @ 19:53, 17 Elul 5771 / 19:53, 16 September 2011 (UTC)

Problem with a design script

Hello -

I had the design script from Wikipedia:User_page_design_center/Scripts successfully installed, and all was fine. I forgot where I was and cleared and saved my vector.js page. I tried to undo my change, but the script not longer works. I reinstalled from Wikipedia:User_page_design_center/Scripts but it still doesn't work. I turned the WikEd gadget on, and it doesn't work either. All was fine this morning until my mistake. I have cleared my cache. What else may I try? Thanks. JMOprof (talk) 16:48, 16 September 2011 (UTC)

Update: Figured it out with the script authors's help. JMOprof (talk) 23:21, 16 September 2011 (UTC)

MediaWiki 1.18 deployment schedule

For our delight and delectation. Summary version: expect things to go a bit funky here on or around October 4th. On the other hand, there are all manner of exciting things included in the release, so buckle your seatbelts! :D Happymelon 23:55, 16 September 2011 (UTC)

Problem reverting image

So File:DSCF1296.JPG was replaced with a less-useful image, but I can't seem to revert it. Any ideas? ▫ JohnnyMrNinja 05:09, 17 September 2011 (UTC)

I get a 404 (not found) error when trying to view the original 2006 version of that image. Maybe this has something to do with your problem. Regards, {|Retro00064|☎talk|✍contribs|} 05:17, 17 September 2011 (UTC).
Several years ago, a database error resulted in the loss of many images; perhaps this is one. Curious, how do you know that the original image was more useful than the current one — are you able to find a thumbnail of the original somewhere? Nyttend (talk) 11:20, 17 September 2011 (UTC)
I've spent the last few days renaming the remaining DSC####.jpg images. When people upload an image they almost always put it into an article, usually in their next edit. This was in the article Racegun until someone uploaded a sunset over it. Now we have no images for that article, and a mediocre sunset picture in its place. So there is no way to get back what was lost in that manner? No back-ups or anything? ▫ JohnnyMrNinja 05:10, 18 September 2011 (UTC)
 

What's going on with the links in the page history that is pictured to the right? Is it simply the presence of an asterisk at the beginning of the page name? Nyttend (talk) 11:55, 17 September 2011 (UTC)

Yes. This is T14974; unfortunately, there is no workaround. Anomie 14:43, 17 September 2011 (UTC)

Editing changes an article's approval ratings!

There's some sort of bug in the system of approval ratings that is currently being used. When an article is edited, its ratings change! Clearly, that is wrong.

For example, the article on the Solar cooker had good ratings a couple of weeks ago. In all categories, its ratings were 4.0 or higher. It must be popular, since many people have rated it, and their ratings are averaged. This afternoon, I made an edit, and after it the ratings in most categories had dropped to 1.0, the lowest possible! For the average to be 1.0, every person must have given it the lowest rating, which can not be true.

I have previously suspected that doing edits changed the ratings, but this is the most dramatic example I have seen.

I'm using IE8, if that is relevant.

DOwenWilliams (talk) 21:51, 17 September 2011 (UTC)

Each rating is discarded once the article has had 30 edits since the time that rating was made. Graham87 00:31, 18 September 2011 (UTC)
Are the displayed numbers of votes correspondingly reduced when votes are discarded? In the Solar Cooker ratings, it says that about 20 people have voted, but since there has been a lot of editing recently maybe only a handful of votes remain to be averaged. DOwenWilliams (talk) 02:34, 18 September 2011 (UTC)
As an experiment, I put in a rating of 5.0 in a category where previously the average rating was shown as 1.0, with 20 votes. The result of my rating was to raise the average to 3.0, so clearly just one vote, not 20, was previously included in the average. Showing the number as 20 was misleading. DOwenWilliams (talk) 02:52, 18 September 2011 (UTC)

Template creating unwanted space

  Resolved

In the Elizabeth of Russia article, there is currently text which reads:

(29 December [O.S. 18 December] 1709–5 January 1762 [O.S. 25 December 1761] ), also

The corresponding code is:

({{OldStyleDate|29 December|1709|18 December}}–{{OldStyleDate|5 January 1762||25 December 1761}}), also

i.e. the problem comes from {{OldStyleDate}}. Is there any way to stop it generating that space after the closing square bracket? Thanks. It Is Me Here t / c 14:32, 15 September 2011 (UTC)

Yeah, but the result would be


(29 December [O.S. 18 December]1709–5 January 1762 [O.S. 25 December 1761])
Choyoołʼįįhí:Seb az86556 > haneʼ 14:59, 15 September 2011 (UTC)
Have a look at {{OldStyleDate/sandbox}} – I believe that my modification, clumsy as it may be, solves the problem. If so, I'll incorporate it into main template. ​—DoRD (talk)​ 15:48, 15 September 2011 (UTC)
(Perhaps someone better-versed in template programming can take a peek at the code as well.) ​—DoRD (talk)​ 16:06, 15 September 2011 (UTC)
I can't say anything about the code, but empirical results are good. It Is Me Here t / c 12:48, 16 September 2011 (UTC)
  Done [3] ​—DoRD (talk)​ 15:09, 16 September 2011 (UTC)
Thanks! It Is Me Here t / c 14:19, 18 September 2011 (UTC)

George Orwell page

At the end of the second paragraph of the Biography section of the George Orwell page (http://en.wikipedia.org/wiki/George_Orwell) is the line "he married no one and was gay." It cannot be seen in the edit view, so I am unable to remove it. I have refreshed the page, and opened the page again in another window and the statement is still there, so it appears bypassing the cache does not eliminate the issue. Just saying. Note: I am using the latest version of Safari. -tlpresn — Preceding unsigned comment added by 99.95.206.185 (talk) 21:47, 17 September 2011 (UTC)

Strange. Cluebot picked it up and reverted, so it should be fine now. Pi.1415926535 (talk) 21:52, 17 September 2011 (UTC)

I've reopened it and refreshed it, and I still see it on the main page but not the edit page. I opened it in Wikibot (which cannot be edited), and see it there as well. No big deal; easy enough to just ignore it. Just letting you know of this anomaly -tlpresn — Preceding unsigned comment added by 99.95.206.185 (talk) 22:03, 17 September 2011 (UTC)

If you see something like that again, try purging the page. Graham87 00:33, 18 September 2011 (UTC)
If you are not logged in, you are not necessarily given the most recent revision of a page when viewing it, so my guess is that you were served this version. However, when you edit a page, whether logged in or not, you are always given the current version, hence the discrepancy. --Redrose64 (talk) 15:18, 18 September 2011 (UTC)

Article appears protected when it isn't

I posted the following to VP/M yesterday (from a different IP address; I use several):

A few minutes ago when I started to write this posting, the article Sunday shopping showed a "view source" link and not an "edit" link, and the mouseover text said it was protected (not just semi-protected). Yet the talk page showed no discussion of this and the edit history showed no recent edit wars.
What puzzled me further was that the edit history also had no entry for a transition to protected status. And now the article no longer shows as protected anyway. So why would it have done so temporarily, for some unknown length of time?

An explanation was posted in response by PrimeHunter:

There is a glitch where unregistered users sometimes see a "View source" link instead of "Edit". If it happens then you can still edit the article by clicking the View source link which has the same url as the Edit link.

Indeed that workaround does the job, although it remains annoying because the "View source" link is only for the whole article and not for sections. In fact, just now it happened to me again with VP/M and I had to open the whole of VP/M for editing in order to post a "Thanks" response to PrimeHunter in the section. (Okay, strictly speaking I didn't have to do it, but it was the least inconvenient way.)

But the reason I'm posting here, of course, is to say that if this is a known problem, then I hope someone will be fixing it!

(Please don't bother suggesting I should register. Not going to happen.)

--70.48.230.233 (talk) 05:16, 18 September 2011 (UTC)

The glitch has been reported here several times, for example at Wikipedia:Village pump (technical)/Archive 88#"Page is protected" - but it is not, and I can edit anyway via View Source in March, and Wikipedia:Village pump (technical)/Archive 91#"View source" on unprotected pages... in June. I haven't found a bug report about it at bugzilla: but it seems time somebody files it. I have very little bugzilla experience and would prefer somebody else. You should be able to fix a given page and get section edit links by purging but that's inconvenient and many users will not know about it. PrimeHunter (talk) 16:05, 18 September 2011 (UTC)
Thanks for pointing out the "purge" workaround and for reopening the bug report at bugzilla. (Original poster, now at...) --142.205.241.254 (talk) 22:13, 3 October 2011 (UTC)

Embedding SVG instead of thumbnailing

It appears that when WP renders an SVG image it actually puts a PNG version (created by RSVG) of the image on the page. I can understand doing that a few years ago when SVG support wasn't universal, but I'm wondering whether that's still a valid reason or if there are other reasons to keep everything in PNG format. I'm thinking at least let people opt in in the user preferences. The reason I'm bringing this up is SVG can be animated and this feature is now supported by the major browsers, and I think animated SVG's have the potential to replace many of the animated GIF's, such as the one in this article, which do not display correctly in a smaller size.--RDBury (talk) 13:32, 18 September 2011 (UTC)

I don't know about IE9, but Internet Exploder 7 certainly cannot handle SVG files. Current versions of Firefox, Google Chrome, Safari and Opera all can, though. --Redrose64 (talk) 15:26, 18 September 2011 (UTC)
According to w3.org, IE 9 has native SVG support while IE 4-8 have it if you install a plug in. Firefox, Opera, Safari, Chrome support most of SVG, and many other browsers have at least some support. Whether or not there is universal support now, support is growing all the time, so it would be good to plan ahead.--RDBury (talk) 20:39, 18 September 2011 (UTC)
There are other reasons than just browser support to generate png thumbnails. The biggest reason is download size. Many of our SVGs are rather complex, and at thumbnail size are often larger in actual bytes than the small PNGs. —TheDJ (talkcontribs) 13:59, 19 September 2011 (UTC)
I'm not convinced that file size is a valid reason, though it's true in some cases that a PNG is smaller than SVG, especially when viewed at thumbnail size. It's not true all the time though, for example the diagram in this article is 5k in PNG at 300 px but the SVG is 2k. The info page offers larger PNG renderings and at 2000px the file size is 51k. If you're worried about mobile devices, w3.org claims "The mobile market appears to like vector graphics, for obvious reasons. For example, it was previously announced that Nintendo's next generation Wii gaming system will include an engine from Opera.com using native SVG 1.1."[4] I think the issue is scaling; if a page is being scaled to a different resolution then is better, in terms of both computation required and quality of the result, to start with a vector based original than to rescale a raster based copy. For those images where the PNG really is smaller, it's perfectly feasible to signal this in the [[image:]] tag, for example use px for size if PNG is preferred and % if SVG is better. I think the real advantage for SVG will come with animation; the size of animated GIF is often measured in megabytes and an equivalent animated SVG may be a few percent of that. We already force animated GIFs to be displayed as a single frame in most cases due to file size, but with file size reduction from a vector based format this would be unnecessary.
It just seems to me that there are advantages to SVG, and we advocate using it here, but many of these advantages are lost when the image is converted to PNG on the page. Also, I think there are people who could create images that take full advantage of features available in SVG but don't because these features are lost when it's filtered to PNG format. In other words give people the option to make diagrams that are interactive or that move and they will take advantage, but it won't happen until the option is available. But I'm not really expecting much from this discussion, though I'm hoping it will be considered as a way that WP can be improved in the future.--RDBury (talk) 15:54, 19 September 2011 (UTC)
Here's one which is 68,785 bytes as a 200px PNG; 406,126 bytes as a 500px PNG; 1,307,005 bytes as a 1000px PNG; 3,585,375 bytes as a 2000px PNG but 8,627,106 bytes as the raw SVG (nominally 1193px wide). I downloaded each of these without previously opening them, then used my browser (Firefox 3.6.22 under Windows XP) to open each one locally, and timed how long it took from clicking the link to the completion of load. All of the PNG versions (including the big 2000px one) loaded so quickly that my stopwatch timings (between 0.5 and 1.2 sec) have a significant margin of error (50% either way?). By contrast, the SVG file took 1 min 52 sec to render, during which time my PC fan, which is temperature sensitive, complained loudly. --Redrose64 (talk) 19:34, 19 September 2011 (UTC)
How consistent are the implementations? Last time I checked (probably about 1.5-2 years ago), none of the platforms that implemented SVG actually implemented the complete SVG standard (except for the plugin from Adobe, I think), and as a result different browsers would often generate appreciably different renderings when given a complex SVG because they varied on which features they supported. We already have problems with some user-created SVGs rendering in unexpected ways when fed to RSVG, and I wouldn't want to amplify that problem by having to support multiple platforms which each have their own quirky rendering differences. Dragons flight (talk) 16:11, 19 September 2011 (UTC)

Downloading Images

Would I get banned/blocked if I ran script that would fetch 800 px wide scaled down image version from image table (imported from dumps/images.sql.gz) at rate 1 per 2 seconds until finished for some 700,000 images (would take around 3-4 weeks). It would be non-intensive non-parallel one-time crawl. Do I need bot account for it or does it qualify as reason to get banned ? Gzilirion (talk) 15:28, 18 September 2011 (UTC)

Is this right place to ask this question ? Gzilirion (talk) 03:26, 19 September 2011 (UTC)

Ops says

yes, it should be fine. I wonder what they want them for... maybe they'd want to provide a tarball of them in batches of 2gb each or something, so that other folks could benefit from the download?

1 request / 2 secs (so 30 requests / 1 minute) should be fine. The most important things are:
  • only do 1 HTTP request at any time, (make them serial not concurrent)
  • use a relevant and unique User-agent header which includes your contact information (email address) so they can contact you if any issues do arise
--Jeremyb (talk) 06:09, 20 September 2011 (UTC)

something wrong with prod tag

Is anyone else having trouble with the prod tag, for some reason its not doing the whole "This message has remained in place for seven days and so the article may be deleted without further notice." message onces its passed the expiry time, like with Stagecoach in Eastbourne bus route 1, its well past the expiry but the template is acting as though it hasn't. However if I open the edit page and show the preview without making any changes, the article will show the expired prod tag as it should. Is anyone else having this issue?--Jac16888 Talk 16:22, 19 September 2011 (UTC)

I've fixed this one by purging the page, but this doesn't solve the underlying problem. Someone/something needs to trigger a purge once the seven-day period has expired. -- John of Reading (talk) 16:36, 19 September 2011 (UTC)
I suppose a purge bot could be programmed but I don't think it's worth it. The article may not show up in Category:Expired proposed deletions but it can still be found in Category:Proposed deletion as of 12 September 2011 so the prod will not remain undiscovered for an indefinite time. Most admins would probably know why the tag hasn't updated, and the tag shows the expiry date, for example at Besst, even if it doesn't say the date has already come. PrimeHunter (talk) 20:45, 19 September 2011 (UTC)

Special:Block not working

I've been trying for about 15 minutes to block an IP vandal, but every time I press "block" (after a few minutes of trying), I get a databse error:

Database error

A database query syntax error has occurred. This may indicate a bug in the software. The last attempted database query was:

(SQL query hidden)

from within function "Block::purgeExpired". Database returned error "1205: Lock wait timeout exceeded; try restarting transaction (10.0.6.46)"

So, er, what does that mean in plain English and when can I block my damn vandal? HJ Mitchell | Penny for your thoughts? 18:01, 19 September 2011 (UTC)

That seems similar to the error message that was occurring when editing talk pages a few days ago. See #Lost access to my talk page. Apparently that was fixed my pinging a dev on IRC, so maybe that's your best bet if this hasn't been fixed yet. Jenks24 (talk) 18:27, 19 September 2011 (UTC)
Lock wait timeout means the software / database didn't complete its operation in the allowed time. This could indicate that the database is simply temporarily overloaded or stuck (it happens sometimes), or it could indicate that the operation that the software is trying to execute is poorly designed and hence takes an unreasonable amount of time. In general, I would ignore timeout errors unless they are frequent and/or reproducible. In most cases, timeouts will fix themselves once database load resolves itself. However, if you get the same error repeatedly, then it is worth filing a bug that describes exactly what you are doing to get there so that other people can investigate. Dragons flight (talk) 18:50, 19 September 2011 (UTC)

Merge

According to this link FurMe is supposed to be merged with Twinkle. Did it happen?Gregory Heffley (talk) 22:55, 19 September 2011 (UTC)

As far as I know, no. — Kudu ~I/O~ 01:00, 21 September 2011 (UTC)
See Wikipedia talk:Twinkle/Archive 24#FurMe news. It has been partially merged and work is apparently ongoing. You can still use the old FurMe by adding to your js. Jenks24 (talk) 01:29, 21 September 2011 (UTC)

Moving a to a title that is currently a redirect

Hey All,

I'd like to move Religious Society of Friends to Quakers. Quakers is currently a redirect to Religious Society of Friends, so simply trying to use the "move" function gives the "article under that title already exists" problem. What's the correct thing to do in this situation? Request a deletion of the redirect, then move? NickCT (talk) 14:22, 20 September 2011 (UTC)

Yes, use {{db-move}}. But I would use the requested moves process for such a high-profile article. Graham87 14:29, 20 September 2011 (UTC)
Thanks! You're probably right about WP:RM. I did try to start a discussion on the article's talk page, but didn't get anywhere.
I'll go down the RM path path..... Don't want to create any Wikicontraversy. Thanks for the info though. It will be helpful in future. NickCT (talk) 14:35, 20 September 2011 (UTC)

Soxred's rangecontribs tool

Soxred's rangecontribs tool is not working for me. I am getting a blank page in Firefox and a HTTP 500 error in IE. Is there any alternative rangecontribs tool available?--Sodabottle (talk) 06:16, 4 September 2011 (UTC)

The tool has been down for a few days now, and has been reported here. I do not believe there are any other similar tools. Sorry. - Hydroxonium (TCV) 01:15, 6 September 2011 (UTC)
There is a gadget in preferences. Plus I have a userscript that supports more ranges than the gadget; I can publish it if anybody wants. — AlexSm 04:00, 6 September 2011 (UTC)
Thanks very much, Alex. Yes, could you please publish the userscript? I know there are a few people here and at X!'s talk page that would appreciate your script. Thanks. - Hydroxonium (TCV) 04:05, 7 September 2011 (UTC)
Sorry for the delay, the script is on the top of this page: user:js, you have to add one line to your common.js before using it. It's still beta but I hope it's useable anyway. — AlexSm 04:27, 13 September 2011 (UTC)

I've rolled my own version of X!'s script; it's available at http://toolserver.org/~helloannyong/range/. — HelloAnnyong (say whaaat?!) 03:25, 17 September 2011 (UTC)

Thank you Alex and thank you HelloAnnyong. It's very generous of both of you to help replace such a needed tool. I'm wondering if we should mention these tools somewhere other than here and X!'s talk page for those who have a need for these useful tools. Just a thought. Thanks again. - Hydroxonium (TCV) 03:17, 22 September 2011 (UTC)

Babel extension

The new Babel extension is scheduled to be deployed on Wednesday, September 21, 2011 (bug 15308) to all Mediawiki projects on an "opt-out" basis. Is there a discussion somewhere at Wikipedia about this extension? I am looking for information to help my home project to decide whether to opt, and to understand the configuration process. Thanks. ~ Ningauble (talk) 17:19, 18 September 2011 (UTC)

More at mw:Extension:Babel. --Redrose64 (talk) 18:30, 18 September 2011 (UTC)
How do I start using it? Any help page?--Kozuch (talk) 18:50, 21 September 2011 (UTC)
Babel user information
en-N This user has a native understanding of English.
fr-1 Cet utilisateur dispose de connaissances de base en français.
Users by language
Yeah, mw:Extension:Babel#Usage. Say that you spoke English like a native, but had a slight amount of French, you would put {{#babel:en-n|fr-1}}, which shows as at right. --Redrose64 (talk) 19:15, 21 September 2011 (UTC)

Smart Phone Editing

Is there a way to get it when i click edit on my phone that it opens in a new window? I've tried 4-5 Android browsers and I don't know if it is my phone or just how it is set up, but I can't scroll down in the edit window on my phone, only down the whole page, so it makes editing impossible. And when I can scroll down, when I try to click on a word to edit, the cursor jumps all over. CTJF83 23:07, 20 September 2011 (UTC)

I don't know what type of phone you have, or wether you've tried Opera mini or not, but I find that it works for me. Nota bene: it only works well for smaller pages or section edits, else it gets overwhlemed and locks up, at least on my phone. (this edit via Android) Avicennasis @ 16:17, 22 Elul 5771 / 16:17, 21 September 2011 (UTC)
I have the htc evo...I tried regular opera and it worked hit or miss...I'll try that one thanks, CTJF83 06:14, 22 September 2011 (UTC)

String function needed

Hi. Is there any functions or template to crop a specified number of characters off the beginning, with at least 200 limit of string length, similar to {{Str crop}}, {{Chop head and tail}}? I want to parse a URL, which may easily exceed 100 characters (I think 200 limit would be fine), http://en.wikipedia.org/w/index.php?title=Blah_blah&action=historysubmit&diff=111111111&oldid=222222222 for example, I want to get the "title", the "diff" ID and the "oldid" of the URL, but there's no string functions installed for Wikipedia. If there's a function to crop off the beginning (200 limit), I can handle this, but I can't find one. --Tomchen1989 (talk) 12:10, 21 September 2011 (UTC)

{{Str right}} and {{str rightc}} will handle input strings of up to 100 chars, but they are expensive. A similar template handling 200 would be even worse. --Redrose64 (talk) 13:31, 21 September 2011 (UTC)
The better question in these kinds of discussions is often: "Why do you need to parse a string like this". There are many types of solutions for many problems, and doing complicated calculations is often not the solution, for most problems you want to solve. 16:10, 21 September 2011 (UTC) — Preceding unsigned comment added by TheDJ (talkcontribs)
Assuming that the question is related to optimizing {{Copied}} parameters, I think that JavaScript or a bot would be a better solution. I considered writing a user script, but the rendered output is functionally the same, and the oldids can't be used for anything like categorization. Flatscan (talk) 04:35, 22 September 2011 (UTC)

Is the filename blacklist not working?

All files named DSC#####.* have been renamed or deleted, and DSC is on the MediaWiki:Filename-prefix-blacklist, so why are there now two new files with that prefix?[5] Shouldn't the blacklist prevent that? ▫ JohnnyMrNinja 18:23, 21 September 2011 (UTC)

This message is only for warnings (at least MediaWiki:Titleblacklist says so). Files names are disallowed with Titleblacklist where you can find the line File:DSC.\d+\.JPG which means DSC...<letters>.JPG is allowed at the moment. — AlexSm 19:42, 21 September 2011 (UTC)

Can't load en.wikipedia.org

Any idea what's going on with the servers? Whenever I try to load any page at en.wikipedia.org (I'm only able to access this page with the secure server right now), I get a message of "Internet Explorer [version 8.0.6001.19120] cannot display the webpage". Nyttend (talk) 05:28, 22 September 2011 (UTC)

Very strange. I have no problem with Chrome nor IE 9.x, logged in or logged out. Is there anything else you're having trouble accessing? Pi.1415926535 (talk) 06:09, 22 September 2011 (UTC)
  Works for me, at least. It might be an issue on your side. mc10 (t/c) 06:32, 22 September 2011 (UTC)
Able to do it now, but for several hours I couldn't access any WMF sites except the secure server. Commons worked occasionally (eventually it started before en:wp did), but normally an attempt to load a Commons category resulted in a vertical list of files (many of which didn't download) with the categories and the "This page was last modified on ", etc. text following. Wikipedia simply produced an error similar to what I'd get if I weren't online at all. Nyttend (talk) 11:21, 22 September 2011 (UTC)
I also experienced some timeouts at a sister-project, and on a least a couple occasions the database was locked from saving edits to allow replication to catch up. For a while, live queries like Special:RecentChanges were returning day-old information. I suspect the early stages of the MediaWiki 1.18 rollout (cf. Signpost coverage) is stressing the servers. I am only guessing, but I would not be surprised to see similar problems during the next couple weeks, until the rollout is complete. ~ Ningauble (talk) 12:43, 22 September 2011 (UTC)
Now I'm back to the same problem as before. I'm at a public building, unlike yesterday when I couldn't access it and this morning when I could, both of which were at my house. Nyttend (talk) 16:42, 22 September 2011 (UTC)
Thanks, I've reported this. Philippe Beaudette, Wikimedia Foundation (talk) 14:14, 23 September 2011 (UTC)
Sorry for the inconvenience folks. We weren't aware of any problems with en.wikipedia.org at the time of these reports. I'm looking into what it might have been, but if it comes up again, please let us know here (or better, one of the channels outlined in the 1.18 blog post). Thanks! -- RobLa-WMF (talk) 17:33, 23 September 2011 (UTC)
Thanks. FYI, the symptoms I described happened at en.wikiquote on September 21. Things have been relatively stable since then, although page loading is sometimes unusually slow. ~ Ningauble (talk) 20:35, 23 September 2011 (UTC)

Using #ifexist with images

Hi all, I have a very large wiki site for literature. I want to upload a photo for each author on the site, but I don't have all the photos. So I thought I could upload the images that I have, and the authors that don't have a photo will get the default image "No photo Available". So I guess I should use the parser #ifexist for that, and tried all kinds of syntax, but none worked.

can anyone help please?

Thank you,

Asaf

MediaWiki 1.16.4

PHP 5.2.17 (cgi-fcgi)

MySQL 5.1.53-log — Preceding unsigned comment added by 80.179.10.221 (talk) 09:57, 22 September 2011 (UTC)

 
ifexist is documented at mw:Help:Extension:ParserFunctions##ifexist. It requires mw:Extension:ParserFunctions. Image details can depend on your site (are the images uploaded locally?). I guess you plan on standardized image names. This may work:
{{#ifexist:File:{{PAGENAME}}.jpg | [[File:{{PAGENAME}}.jpg|100px|right]] | [[File:Image is needed male.svg|100px|right]]}}
It displays File:Raymond Queneau.jpg if it's previewed at Raymond Queneau, and File:Replace this image male.svg on most other pages, for example this one. If it doesn't work then please post a link to a page at your site where the author has an image. PrimeHunter (talk) 11:30, 22 September 2011 (UTC)
Use the media prefix instead of file, and it will work with non-local files too. ΔT The only constant 11:32, 22 September 2011 (UTC)

Math problem: page views by topic

I've got a user (on MediaWiki) who has decided that articles about entertainment professionals account for half of the server traffic.

Does anyone have a way to estimate traffic by general subject category? That is, not just "There are about a million articles about dead and living people, out of nearly four million articles on the whole site", but "The article Freddie Mercury gets more than 110,000 hits per day, plus Adele (singer) gets almost 41,000, plus (insert long list of "celebrities"), out of X billion articles served each day, so articles about celebrities account for ___% of traffic". WhatamIdoing (talk) 17:32, 22 September 2011 (UTC)

tools:~alexz/pop and more specifically tools:~alexz/pop/view.php seem to be somewhat what you're looking for. It goes by wikiproject and only those who have opted in. Also note that Mr.Z-man seems to be aggregating hits from redirects (so hits on redirects will be shown as a hit on their target page in his tool's results). Killiondude (talk) 00:46, 24 September 2011 (UTC)

Results from the Huggle experiment

Hi everyone,

This note is to follow up on a past experiment we ran with Huggle and its level 1 warnings. We A/B tested eight templates, including the current default, and the rest with some variation of personalization (e.g. "I reverted you because", "my talk page"), teaching (e.g. "please use the sandbox in the future"), and whether there were images. The full set we tested is here.

There were 3241 total instances of the templates were delivered during the experiment. Some of the general results we found were as follows:

  1. 63% of the people warned stopped editing for the immediate future. How much editing a person did before they were warned was really the determining factor in whether they kept editing afterwards. But for those who did stay after being warned, the type of message made a big difference for how they behaved after.
  2. There was no measurable difference between the templates that had an image and those that didn't. We combined those two groups for the remainder of our analysis.
  3. Only about half of people actually read their talk page messages. We separated out the 1750 who actually read their messages during the experiment and further analyzed outcomes for that group only.

We measured the following different outcomes individually and in aggregate: Contact between the editor and the warning Huggler in any space and whether it was positive, "stays" (or retention measured by continued editing), and "improves" (or cases where the quality of edits improved after being warned).

  • Improvements in editing: The only clear predictor of whether someone improved was their editing history prior to being warned. None of the templates made a statistically significant difference there.
  • Retention: Again, the really significant predictor here was prior editing history, especially if someone had participated in talk pages prior to being warned. However, it was clear that...
    • Adding personalization and teaching material in one message had a negative impact on retention. This is likely because it's longer and may overwhelm people. It probably feels like a lecture.
    • There was a marginal significance that individually, the teaching and personalized messages both lead to more editors staying compared to the current default warning. This is good, but not strong enough statistical significance for us to declare a victory for improving editor retention.
    • Teaching messages alone made those that were clearly blatant vandals leave (at least temporarily) at a higher rate than the current template. See the caveat to this in Contact though.
  • Contact between Hugglers and those warned: We separated out both positive and negative contact between Hugglers warning people and the editors being warned. The conclusions were...
    • For editors who were clearly blatant vandals, we had the unexpected result that the personalized template lead to a decreased likelihood of negative contact (AKA retaliation) than the other templates.
    • Personalized templates lead to editors who weren't blatant vandals contacting the Huggler at a higher rate with constructive things to say and genuine questions.
    • Teaching messages that lacked personalization lead to a higher likelihood of negative contact (including compared to the current default level 1).

The complete results are on Meta. But if our original hypotheses was whether we can change editor behavior by tweaking the content of warnings, the answer is definitely yes.

The unanswered question is: can tweaking the level and kind of personalization lead to more retention of good faith editors, and can we really encourage them to improve? It's clear so far that templates which lack personalization and try to teach editors more can have nasty outcomes like increased retaliation, so we don't want to go down that path.

We've worked up three variations on the last set that we'd like to test further. You can read them at {{Uw-vandal-rand1}}, but the basic lowdown is that we'd like to test just the default, the personalized one, and a third variation which lacks directive speech. The last one is interesting to us because based on the results of the teaching template and some other anecdotal evidence we've seen, we hypothesize that editors don't actually learn what they need to if you use language that sounds like you're assigning them work. Steven Walling (WMF) • talk 19:37, 22 September 2011 (UTC)

With the templates, we should consider making your own. Then, you could say better and more. ~~Ebe123~~ (+) talk
Contribs
21:56, 22 September 2011 (UTC)
I absolutely agree that custom-tailored responses are always superior to templates. But there is a very large volume of responses to give, so I also understand the appeal of the expediency that templates provide. Steven Walling (WMF) • talk 21:07, 23 September 2011 (UTC)

Problem with template infobox professor

I just tried to add an infobox professor to Cornelis de Jager. I did not get the desired infobox, but the definition of the infobox. I quickly reverted the change. Could someone look what went wrong? Pimvantend (talk) 16:35, 23 September 2011 (UTC)

There isn't any such infobox anymore, per Wikipedia:Templates for discussion/Log/2011 May 30. The page was re-created 26 July 2011, but contains nothing but the documentation for the nonexistent template. Weird. Ntsimp (talk) 17:03, 23 September 2011 (UTC)
I've tagged it for speedy delete per WP:CSD#G4. --Redrose64 (talk) 17:20, 23 September 2011 (UTC)

Seeking help with template format

I noticed that Template:Circular-ref displays like this: [circular reference]

with that extraneous bracketing at the end. (Posting this message here, I also noticed that it doesn't close the small italic font formatting unless there's a line break after it.) Could someone who has more skill with templates than I do please clean that up? Thanks. --Tryptofish (talk) 18:11, 23 September 2011 (UTC)

FixedDrilnoth (T/C) 18:24, 23 September 2011 (UTC)
Outstanding! Thanks! --Tryptofish (talk) 18:26, 23 September 2011 (UTC)

I hate Wikipedia's servers

This is the only popular website that doesn't always come up immediately when clicked (via any link, including internal ones). Sometimes a second click helps, and will cause an instant load. It's potluck. Some clicks just seem to hang and go nowhere. I'm not sure if I've ever been patient enough to wait and see if they eventually load.

The problem has been getting worse over the years. I noticed the progression a lot more after taking a long Wikipedia break and then coming back. And don't get me started on how much edit preview load times have crept up in proportion.

It's not my ISP and it's not my computer. I'm a techie by profession and have run through the standard script already. I'm not complaining about anyone and I'm not trying to fling crap. It's just getting to be a frustrating situation. I know everyone involved works hard and I appreciate it. I've brought this up before and feel it should be looked into again. That's all. Thanks. Equazcion (talk) 20:35, 23 Sep 2011 (UTC)

Well, I (from Germany) can't reproduce that at all... may you want to give us some further information (location (country should be enough), logged in or logged out (logged out viewed pages are served differently than logged in ones), articles or special pages?) - Hoo man (talk) 20:59, 23 September 2011 (UTC)
Usually when people are complaining that Wikipedia is awfully slow all the time (even for reading), it is because your personal preferences are set in an unusual manner that forces the pages to rerendered on nearly every load. Using the Stub Threshold or having a thumbnail size other than the default (220px) are big offenders for this. For a simple test, you can try creating a new account and see if reading and editing seem to move more quickly. If so, then it is almost certainly related to your preferences. Similarly you can try loading pages while logged out, and see if they appear to load much more rapidly. Editing is sometimes slow but reading pages should usually move quickly. Dragons flight (talk) 21:11, 23 September 2011 (UTC)
I thought - and someone can correct me if wrong - that pages weren't nearly as cached for logged-in users anyway, and thus logging out would speed up any setup without necessarily showing that personal preferences were influencing that. Grandiose (me, talk, contribs) 21:47, 23 September 2011 (UTC)
Indeed so. The anonymous users all get the same version of every page and this has the best caching of everything on Wikipedia. If you are logged in, what helps might be the following: Reset your preferences to the defaults. The caching for logged in users is fragmented per page by the set of preferences that you use. Using the defaults means having the largest chance of having the page already cached. —TheDJ (talkcontribs) 17:05, 24 September 2011 (UTC)

Is There A Source Download Available For The List of Blocked IP's.

I am looking to access a list of blocked IP addresses which have been blocked due to being associated with web proxies and other such anonymity services which could be used to bypass the security of my site. Does Wikipedia have this list available which can be downloaded regularly from the site for integration with an external database? — Preceding unsigned comment added by Knightcon (talkcontribs) 03:07, 24 September 2011 (UTC)

Wikipedia:Database reports/Range blocks is a list of IP ranges that are blocked at Wikipedia. Many of these are listed as proxies in the notes section of the report. Could this be what you're looking for? - Hydroxonium (TCV) 03:34, 24 September 2011 (UTC)
There's no central list. Wikipedia:Database reports/Indefinitely blocked IPs gives some more, there's a list of nearly 30,000 at CAT:OP, User:ProcseeBot has made nearly half a million proxy blocks[6], TorNodeBot has made over 10,000 blocks[7], RonaldB has also blocked a shedload with his bot[8], and the global IP block log contains another load of abusive IPs, including many open proxies.[9]. In my opinion this is not a good approach. There's simply too many, they're always changing, and the collateral would be huge. You would be better using something which checks someone else's database or DNSBL. Project Honey Pot and Tor (anonymity network) both provide some API you might be interested in. Or even better, just make sure your security is solid. -- zzuuzz (talk) 10:37, 24 September 2011 (UTC)

Signature Problems

Hello. I was wondering if someone could assist me with a minor problem I am having in regards to my account signature. As you can clearly see from this edit, my signature appears to link fine... Here is an example of where the signature works fine on this page as well: Stubbleboy. Problem is, when I add the signature to my preferences, and then attempt to use the signature with tides there, it shows up as seen on my user page. I just can't make heads or tails of why I can't get the link to work. Also, the "Stubble" portion was the color "scarlet", however now it appears that is not working as well? The formula for the signature I am using is [[User:Stubbleboy|<b style="color:scarlet">Stubble</b>]][[User talk:Stubbleboy|<b style="color:gray">boy</b>]]. Thanks to anyone who has time to help me out. Stubbleboy 21:05, 24 September 2011 (UTC)

When any page contains a link to itself, that link will appear as boldface black. For example, putting [[Wikipedia:Village pump (technical)]] on this page shows as Wikipedia:Village pump (technical). --Redrose64 (talk) 21:09, 24 September 2011 (UTC)
I see, and as for the color issue? Stubbleboy 21:20, 24 September 2011 (UTC)
Well, may have answered my own question. Perhaps the color "scarlet" is non-existent because when I changed it to "red" I received a different result... Stubbleboy 21:22, 24 September 2011 (UTC)
There are only 16 officially defined color names for HTML. CSS3 defines a rather larger set, but still no scarlet. Personally, I'd stick with RGB codes or the 16 HTML color names. Anomie 21:55, 24 September 2011 (UTC)
(edit conflict) See here where there is no mention of "scarlet". Note that some browsers don't support all of these names: you're quite safe with "red" though. --Redrose64 (talk) 21:57, 24 September 2011 (UTC)
Appreciate both of your assistance. Anomie, best of luck in your RFA. Stubbleboy 22:09, 24 September 2011 (UTC)

New template requested — Discreet abbreviation

Hello,

See this discussion.

--Nnemo (talk) 23:13, 24 September 2011 (UTC)

Edit summary max length and behaviour

Hello,

I encounter an annoyance and a bug in the English Wikipedia.

The edit summary has a max length a bit too short.

I am typing some text, and I reach the max length. So I select a few characters, and I type the letter "y". This must replace the characters by a "y". This does not. I guess that the culprit is some crappy ayatollesque JavaScript input restriction.

Please improve and correct that. Thanks.

--Nnemo (talk) 21:01, 24 September 2011 (UTC)

I can confirm this in Firefox 6.0.2 and Chrome 14.0.835.186. Ucucha (talk) 21:04, 24 September 2011 (UTC)
The maximum length was increased in February 2011 and is unlikely to be increased again. In cases as described above, the workaround is to select the text, then press either "Delete" or the backspace key; and then type in your new text. --Redrose64 (talk) 21:11, 24 September 2011 (UTC)
Yet why should that workaround be necessary? To be clear, this bug happens when you are at the maximum edit summary length and try to replace a series of characters in the input box with a single character (so you are not going past the maximum length). That ought to be possible. Ucucha (talk) 21:55, 24 September 2011 (UTC)
Without access to the low-level source I can't be certain, but I might guess that it tries to insert the new text before the old is actually deleted. This is probably a question to ask at MediaWiki, who deal with the user interface. --Redrose64 (talk) 22:01, 24 September 2011 (UTC)
I suppose so, and in that case the problem is more likely to be in the browser than in MediaWiki. Ucucha (talk) 22:02, 24 September 2011 (UTC)
Oh no, the problem does not lies in the Web browser, but in the Web site. I had first believed my Web browser was wrong, so I tested him, and he is perfectly innocent.
--Nnemo (talk) 22:41, 24 September 2011 (UTC)
I'm sorry for the incorrect assumption. Ucucha (talk) 12:46, 25 September 2011 (UTC)
The code is available at http://svn.wikimedia.org/viewvc/mediawiki/trunk/phase3/resources/jquery/jquery.byteLimit.js?view=co. The problem is that it doesn't take into account the length of the selected text that will be deleted when checking to see if the new character will fit. It appears that determining the selected range must be done differently on different browsers.
FYI, the maximum length is 250 bytes (not characters) because the underlying database field has a maximum length of 255 bytes; the limit on the edit summary field was formerly lower because an HTML input field's maxlength property works in characters and no javascript was being used to determine the byte length of the characters entered, so it had to be pessimistic. Anomie 04:59, 25 September 2011 (UTC)
I can confirm this is the intented behaviour: once you reach the limit only special keys (like del or backspace) are accepted. The source that's live now is probably this one; later it was moved to jquery.byteLimit.js (not live yet). I suspect that the code to allow you to paste-then-undo would have to be much more complicated. — AlexSm 04:48, 25 September 2011 (UTC)
This is bug #29467. You can vote for it if you have an account on Bugzilla. Helder 22:26, 25 September 2011 (UTC)

Preferences → Gadgets → Allow up to 50 more characters in each of your edit summaries. ---— Gadget850 (Ed) talk 12:13, 25 September 2011 (UTC)

I've noticed that strange setting. Is there a brief explanation somewhere? How can ticking a box provide room for 50 more characters? Johnuniq (talk) 12:22, 25 September 2011 (UTC)
I think it has to do with what Anomie said: the hard (database) limit on the size of the field is 255 bytes, but the HTML used to limit the size of the edit summary measured in characters, which may be more than one byte long. Therefore, the standard maximum length of the edit summary is set at 200 instead of 255 to allow some room for multibyte characters. However, with JavaScript it is possible to check not just the character but also the byte count, so it is safe to go up to 250 characters. Ucucha (talk) 12:46, 25 September 2011 (UTC)
Yes. I dug up the discussion at Wikipedia:Gadget/proposals/Archive 2#Long edit summaries.
That gadget has had no effect since the rollout of MediaWiki 1.17 in February 2011 - whether it's set "on" or "off", the limit is always 250 chars. --Redrose64 (talk) 13:18, 25 September 2011 (UTC)

Who can fix a table?

I am reworking some lists and would like to have two tables on the page look alike. The top one, at User:GeorgeLouis/Sandbox3#1889.E2.80.931909_.28nine_wards.29, should look like the bottom one, at User:GeorgeLouis/Sandbox3#1925_and_after_.28fifteen_districts.29. In other words the "wards" table should have the names aligned at the top of the columns instead of being repeated, as they are now. Who can assist? Beseechingly, I am GeorgeLouis (talk) 07:52, 25 September 2011 (UTC)

  Done -- John of Reading (talk) 10:51, 25 September 2011 (UTC)
  Resolved

Hi, in this project page the "Edit section" link doesn't appear as it should. Can't figure out why. Also the TOC didn't appear automatically as it should then I forced it with the "_TOC_". Anyone knows why is this happening? --Tachfin (talk) 11:42, 25 September 2011 (UTC)

Wikipedia:WikiProject Morocco/Tabs includes __NOTOC__. There is probably a template that includes __NOEDITSECTION__ ---— Gadget850 (Ed) talk 11:59, 25 September 2011 (UTC)
The problem seems to be in the templates placed at the top in fact. When I remove them the "edit section" link appears. I suspected that before and tried to solve the problem with <noinclude> but didn't work. I verified the templates they don't include the _NOEDITSECTION_ Tachfin (talk) 12:23, 25 September 2011 (UTC)
Special:ExpandTemplates is your friend. It turns out that the project page is transcluding Template:Box-header about four levels deep, which puts __NOTOC__ and __NOEDITSECTION__ on the page. I'll try to fix it. Ucucha (talk) 12:51, 25 September 2011 (UTC)
Yes that seems to be the source of the problem. From template:Box-header, what does this do "{{#if: {{{TOC|}}} | |__NOTOC__}}{{#if: {{{EDIT|}}} | |__NOEDITSECTION__}}"? Tachfin (talk) 13:02, 25 September 2011 (UTC)
It sets __NOTOC__ if the |TOC= parameter is not set, and does the analogous thing for |edit=. Ucucha (talk) 13:05, 25 September 2011 (UTC)
Your edit worked I restored them. I changed the template transcluded on projects page before I guess that's why you reverted your edit...Thank you! Tachfin (talk) 13:17, 25 September 2011 (UTC)

bits.wikimedia.org

In particular http://bits.wikimedia.org/meta.wikimedia.org/load.php?debug=false&lang=en&modules=jquery%7Cmediawiki&only=scripts&skin=monobook&version=20110912T172820Z:1 - this page (or bits in general) takes a load of downloading (32k, apparently with every WP page) and often stalls the browser trying to execute. What's it supposed to be doing, and how can I turn it off. (Incidentally I'm not terrifically happy with cross-site scripting, even if it is in the same second level domain.) Rich Farmbrough, 15:21, 25 September 2011 (UTC).

It's all the javascript for jQuery, the core mediawiki script loader, and the scripts for the monobook skin. It's not downloading every page request, because your browser will cache it (unless your browser is stupid/broken). It is the basis for all javascript execution on wikipedia pages. It cannot be disabled, other than by fully disabling Javascript for the entire website I think. The reason it is on a different server is actually that it has an entirely different caching architecture. —TheDJ (talkcontribs) 16:17, 25 September 2011 (UTC)
(The caching is Varnish rather than Squid (I believe) for anyone interested.) It may be two distinct issues, then. One is that the script is not "responding" according to Firefox, basically it's some sort of resource hog. Secondly I certainly see "waiting for bits.wikimedia.org" regularly when I open a page. Rich Farmbrough, 23:16, 25 September 2011 (UTC).
Must say I get the same thing and the loading stops for ages at that point. Keith D (talk) 23:54, 25 September 2011 (UTC)
Digging a little more, I notice a couple of anomalies. Firstly it also calls for the Chick/main.css. Secondly each of the connections (5 to bits, 2 to text and 1 to wikimedia-lb (all at esams.wikimedia.org)) was closed with a RST, which seems most odd. Rich Farmbrough, 01:27, 26 September 2011 (UTC).

Javascript help

I have tried to narrow down the problem, to no avail. Just to note, I am not that experienced with Javascript, but I cannot for the life of me tell what is wrong with my addtool.js script. Any help with the javascript code I have used would be greatly appreciated.

P.s. I think I may know one or two possible reasons for the problem, but I can't exactly figure it out. LikeLakers2 (talk | Sign my guestbook!) 22:50, 25 September 2011 (UTC)

Could you tell us what the script is supposed to do and what is going wrong? That makes it a lot easier to debug. In looking over your code, it seems as if you're using rather more code, especially more functions, than you really need. It looks like you could do pretty much all the processing you do with a few switches and variable assignments. Ucucha (talk) 23:28, 25 September 2011 (UTC)
It is basically a easy toolbox tool adder. This code:
addtool("tb","javascript:Asteroids()","Play Asteroids2","ca-testtool-tb","Test tool");
Is supposed to work the same as:
addPortletLink("p-tb", "javascript:Asteroids()", "Play Asteroids2", "ca-testtool-tb", "Test tool", "");
But apparently, something in the code is making MediaWiki not want to do that. I don't know what it is that is causing it. LikeLakers2 (talk | Sign my guestbook!) 23:39, 25 September 2011 (UTC)
Also, the functions are basically to split it into parts. LikeLakers2 (talk | Sign my guestbook!) 23:39, 25 September 2011 (UTC)
Sometimes splitting code into parts isn't a good idea—here, it makes the code only less readable to me.
Running this code in Firebug tells me that "addtool2 is not defined". So, the reason your code isn't executing is presumably that you're using an object "addtool2" that doesn't exist. Ucucha (talk) 23:52, 25 September 2011 (UTC)
So it would have been better as step2(), step3(), etc.? That seems to be what you are saying. LikeLakers2 (talk | Sign my guestbook!) 23:58, 25 September 2011 (UTC)
That might make it work, yes. It wouldn't be very good coding, though. Ucucha (talk) 00:02, 26 September 2011 (UTC)
Having undone the edit I thought was related to what you said, as seen here, it still doesn't seem to work, apparently. Mind rerunning it through FireBug, but with the current code? LikeLakers2 (talk | Sign my guestbook!) 00:04, 26 September 2011 (UTC)
Firebug doesn't return any errors, which means it parses correctly. However, when I make it execute addtool(), it still complains about addtool2 not being defined; I guess I have it cached somewhere. Ucucha (talk) 00:23, 26 September 2011 (UTC)
After figuring out that Internet Explorer already has a javascript debugger, I've finally figured it out. I apparently need to wrap the addtool functions around in a onloadhook function, like so:
addOnloadHook(function(){
addtool("tb","javascript:Asteroids()","Play Asteroids2","ca-testtool-tb","Test tool");
});
That apparently makes it work. LikeLakers2 (talk | Sign my guestbook!) 00:26, 26 September 2011 (UTC)

Creating pages with preloaded content

Hi. I asked this at the Help desk yesterday but have gotten no response, so I thought I would ask here. If I need to ask somewhere else, please direct me.

I would like to be able to create a page from a redlink with preloaded content. this section on WP:RFA does it with an input box using the |preload=page parameter, but I would like to do it with a redlink. There are redlinks in Template:NRHP Article Archive and Template:NRHP Picture Archive that users click on to archive new pictures/articles every month. Each month's archive has the same basic formatting: archive header at the top, followed by a search box, followed by a link to the opposite archive (i.e. picture archives link to article archives and vice versa), then an h2 heading, followed by the list. I've successfully moved a lot of this into the two templates above, but when a user clicks on the redlinks, I would like for the resultant pages to preload the template call, as well as the month/year heading seen on all the archive pages in the above templates. Specifically, I would like the new pages to be preloaded with the following content:

Article archives
{{NRHP Article Archive}}
Picture archives
{{NRHP Picture Archive}}
AFAIK, it is not (strictly speaking) possible to preload with a redlink. But you could do some fancy template stuff to make it happen if you really want to. — This, that, and the other (talk) 10:14, 26 September 2011 (UTC)

Language Select

Why is Language select not working? Is it disabled? ~~Ebe123~~ (+) talk
Contribs
10:44, 23 September 2011 (UTC)

Are you referring to the Language field at Special:Preferences? Please specify which language you try to select, whether you do it there and click Save, and where it is not working. I saw many messages change when I tried German and Danish at Special:Preferences but not everything is translated. PrimeHunter (talk) 17:58, 23 September 2011 (UTC)
The language selection in Preferences has a large number of selections, but defaults to en. Changing the language does not translate pages; for example, setting en-gb for British English does not change color to colour.
The language selection allows localization for messages. Many system messages that appear in the Wikipedia interface are stored in the MediaWiki namespace. Each of those pages has a version for each language selection. Many of those interface pages do not exist by default or have been heavily customized for the English Wikipedia. Only a very few of the non-en pages have been customized. You can view system messages through Special:AllMessages and select the desired language.
Frankly, I don't see the language selection as very useful. Even if all messages were translated, the article content would be untouched. ---— Gadget850 (Ed) talk 19:09, 23 September 2011 (UTC)
I'm talking about:
<div class="multilingual">
<div class="lang-en" lang="en">...</div>
<div class="lang-fr" lang="fr">...</div>
</div>

. It is not working. My selected language is English. ~~Ebe123~~ (+) talk
Contribs
20:35, 23 September 2011 (UTC)

This refers to a feature at meta:Meta:Language select. I haven't heard of it before and don't know how it works. I'm not sure of the point in using it at the English Wikipedia or other wikis for a specific language. PrimeHunter (talk) 22:33, 23 September 2011 (UTC)
This is implemented in commons:Template:LangSwitch, for instance. It's not useful here. – Adrignola talk 00:15, 24 September 2011 (UTC)
I know, but it could be useful here too. ~~Ebe123~~ (+) talk
Contribs
10:06, 24 September 2011 (UTC)
Go to http://te.wikipedia.org/wiki/ప్రత్యేక:గుంపుహక్కులజాబితా and tell me whether Bureaucrats on the Telugu Wikipedia are able to remove administrator rights (a perfectly reasonable question). Now go to http://te.wikipedia.org/wiki/ప్రత్యేక:గుంపుహక్కులజాబితా?uselang=en and tell me the same thing. That's what the language selector is for. Happymelon 11:11, 24 September 2011 (UTC)
Your Telugu links are for an automatically generated special page called Special:ListGroupRights here. A chosen language for special pages and interface messages can be useful but does the displayed language in these have anything to do with the feature Ebe123 is asking for? I thought Ebe123's feature request was for text written by editors on normal pages. I think editors will rarely bother to write text in multiple languages at a Wikipedia for a specific language, and if the feature exists then there are potential problems such as language-specific vandalism or gross errors that may never be discovered because it's never seen by an editor who knows the language and bothers to fix it. Even with well-meaning editors, it also seems a big work to maintain a page in multiple languages and keep the language versions in sync. PrimeHunter (talk) 13:39, 24 September 2011 (UTC)
That is why we have Wikipedia in multiple languages, and the interlanguage link feature. If I want to write in English about the city of Oxford, I do it at en:Oxford; if I want to write in French, I do it at fr:Oxford. When editing that French page, the stuff around the edit box is all in French, unless I set my language to English, when some of it gets translated - including, crucially, the "Save page", etc. buttons. These are not always in the same places as on English Wikipedia; the Chinese Wikipedia, for instance, doesn't present the "Save page" button until you have already tried "Show preview". --Redrose64 (talk) 18:07, 24 September 2011 (UTC)
Precisely; why would you want the English wikipedia to host content in a language other than english? We even have a CSD criterion for such material, IIRC. The interface language selector exists to make it easier to work and edit on wikis in languages that are not your native tongue. Happymelon 17:51, 24 September 2011 (UTC)

In fairness to the OP, it's more a matter of "that's not how we do that" than "it's not useful/a good idea". Indeed there are pages on en: where it would be useful, I'm thinking, for example, of the "please do not post content in Klingon without translating them first" templates, and maybe pages around intertanslatewiki. Rich Farmbrough, 22:19, 26 September 2011 (UTC).

Broken Flag Box

Can anyone determine what is the bug when you put the Northern Mariana Islands into this user box?

 This user has visited Northern Mariana Islands.

There is a flag icon present, as an SVG but the box doesnt link to it. This is the only region I've had trouble creating the userbox. Thank you! -OberRanks (talk) 04:37, 26 September 2011 (UTC)

The flag SVG is File:Flag of the Northern Mariana Islands.svg; see the following. It uses a redirect, but the flag shows. --Stemonitis (talk) 07:48, 26 September 2011 (UTC)
 This user has visited the Northern Mariana Islands.
I have added the flag name to the userbox [10] so both versions display the flag now. PrimeHunter (talk) 10:18, 26 September 2011 (UTC)
Thank you very much! I was starting up a Travel Log for my user page and this was the last box to add. It looks great! -OberRanks (talk) 16:09, 26 September 2011 (UTC)

User talk:168.213.7.230

As I often do on school IP pages with a history of vandalism, I attempted to collapse the bulk of the previous warnings and block notices as they clutter the page. The idea being that students will see the notice to create an account at home and use it to edit at school. Anyway, for some reason in this instance it seems to have made a complete copy of all the notices and placed them inside the collapsed area, adding to the clutter instead of reducing it. What's up with that and how can it be fixed? Beeblebrox (talk) 16:51, 26 September 2011 (UTC)

  Fixed, see here, also Wikipedia:WikiProject user warnings/Usage and layout#Archiving warnings for anonymous users. --Redrose64 (talk) 17:49, 26 September 2011 (UTC)
Oh, cool. Actually I've seen that before but never used it myself. Still curious as to why normal collapsing didn't work, but thanks for the fix! Beeblebrox (talk) 17:57, 26 September 2011 (UTC)
I think that there is an imbalance between opening <div> and closing </div>. Kralizec! (talk · contribs) issued a {{subst:uw-vblock}} at 16:18, 4 November 2008, and that template begins with an opening <div class="user-block"> then a Image:Stop x nuvola with clock.svg; if I add a second opening <div> immediately before those, your version seems to work. To be certain I would need to analyse the page properly; ideally I would then check the various warning templates against their originals to see if one of them has a <div>...</div> imbalance. --Redrose64 (talk) 19:27, 26 September 2011 (UTC)

Weird new look?

The editing box just changed its look-the buttons are now above where the used to be and I don't have my special toolbar anymore. Also, the search box's font is larger and changed. What just happened? Also, the toolbar functionality is horrible and I don't have a reference button anymore. How do I switch back? HurricaneFan25 17:07, 26 September 2011 (UTC)

Me too. Cite and link options vanished in the blink of an eye. As I type this, the "Page notice" at the top repeats itself. I have Firefox 3.6.22 with Windows XP. Maile66 (talk) 17:16, 26 September 2011 (UTC)
Also, on Google Chrome, there are no 'show' tags in the WikiProject boxes... – Plarem (User talk contribs) 17:25, 26 September 2011 (UTC)
Me also (Firefox). I have the clock gadget for refreshing pages, and that appears only on my watchlist but not anywhere I want to edit. I suspect some incomplete software update is happening, and isn't finished, so that some functionality is lost. This happened a few days ago too, then went back to normal. --Tryptofish (talk) 17:29, 26 September 2011 (UTC)
For me, it went back to normal just as I saved the comment above. --Tryptofish (talk) 17:30, 26 September 2011 (UTC)
Mine just went back to normal, too. System must have hiccuped. Maile66 (talk) 17:31, 26 September 2011 (UTC)
Mine still looks like the time I changed it to monobook. I am using Firefox 3 on a MacBook Air. HurricaneFan25 17:33, 26 September 2011 (UTC)

See #What's going on with the servers?, below, for what is probably the answer. --Tryptofish (talk) 00:04, 27 September 2011 (UTC)

Wikipedia is a phising site?

So when I try and log in at English Wikipedia, I get this message in the log in screen:

You are viewing this page on //en.wikipedia.org, which might be a proxy or phishing site. This site can intercept your password; you are strongly advised to log in from en.wikipedia.org.

So, I logged in at Meta. Is this something new or a problem with me? My IP hasn't changed (this didn't occur yesterday or when I logged in my alternate account earlier today) and I haven't reconfigured any settings. Elockid (Talk) 20:17, 26 September 2011 (UTC)

What browser are you using? I wonder whether this might be related in some way to protocol-relative URLs. Ucucha (talk) 20:23, 26 September 2011 (UTC)
I tried it on Firefox (6.0.2) and Chrome (14.0.835.186). Elockid (Talk) 20:26, 26 September 2011 (UTC)
I am experiencing the same issue with Firefox 6.0.2. The secure site login is fine. mc10 (t/c) 20:31, 26 September 2011 (UTC)
I experienced this same issue. I was logging in from the login screen, and got this message in a red box:

You are viewing this page on //en.wikipedia.org, which might be a proxy or phishing site. This site can intercept your password; you are strongly advised to log in from en.wikipedia.org.

-download ׀ talk 20:29, 26 September 2011 (UTC)
Was logging in from [11] with Firefox 6.0. -download ׀ talk 20:30, 26 September 2011 (UTC)
See the thread directly above this. Killiondude (talk) 20:31, 26 September 2011 (UTC)
I get the same message. It's apparently made by MediaWiki:Loginend because {{SERVERNAME}} currently returns //en.wikipedia.org and not en.wikipedia.org. At the time this page is rendered, {{SERVERNAME}} returns en.wikipedia.org. PrimeHunter (talk) 20:32, 26 September 2011 (UTC)
I see. Thanks for the info. So, nothing to worry about then? Elockid (Talk) 20:38, 26 September 2011 (UTC)
No. This edit by WOSlinker (I was about to do the same thing) should fix the issue. Ucucha (talk) 20:40, 26 September 2011 (UTC)
This may be related to the impending MediaWiki 1.18 rollout. One of the new features is that omitting the http: from a full URL will detect whether you're coming from the normal server or the secure server, and adjust the URL accordingly. See mw:MediaWiki 1.18#Protocol-relative URLs. --Redrose64 (talk) 20:40, 26 September 2011 (UTC)
Thanks guys for the quick responses. Elockid (Talk) 20:41, 26 September 2011 (UTC)
Confirm it is fixed now. Thanks all. Geoff Who, me? 21:29, 26 September 2011 (UTC)
Reported in bugzilla:31176. 62.140.137.139 (talk) 23:14, 26 September 2011 (UTC)
{{SERVERNAME}} has been fixed. It returns en.wikipedia.org at the time of writing, as it should. It returns en.wikipedia.org at the time of rendering.
But I think there is still an issue with {{SERVER}}. At the time of writing it returns //en.wikipedia.org. At the time of rendering it returns //en.wikipedia.org. As far as I know it returned http://en.wikipedia.org before yesterday and is supposed to return that. I haven't found a way to search for instances of "{{SERVER}}" so I don't know whether this change affects the English Wikipedia. If you think you know a search method that would catch {{SERVER}} but it doesn't find any occurrence then try whether the method can find the old known occurrence of {{SERVERNAME}} at MediaWiki:Loginend. PrimeHunter (talk) 14:29, 27 September 2011 (UTC)
You should read the first comment of the bug report. "{{SERVER}} which is supposed to output the protocol and host)" the protocol is now protocol relative, so correctly returning //en.wikipedia.org, which indeed is different from before, nevertheless correct and expected behavior. —TheDJ (talkcontribs) 22:28, 27 September 2011 (UTC)
Thanks for the explanation. PrimeHunter (talk) 23:59, 27 September 2011 (UTC)

What's going on with the servers?

I keep getting messages today that the servers are having technical problems. The info at the bottom of the message is of the following sort (from the most recent time I got this):

Request: GET http://en.wikipedia.org/wiki/Wikipedia:Village_pump_(technical), from 208.80.152.71 via sq65.wikimedia.org (squid/2.7.STABLE9) to () Error: ERR_CANNOT_FORWARD, errno (11) Resource temporarily unavailable at Mon, 26 Sep 2011 22:48:41 GMT

What's going on? LadyofShalott 22:54, 26 September 2011 (UTC)

They are preparing the databases for the 1.18 upgrade. For this they need to switch around something with the master databases. Two of these switches failed and took everything down for a few minutes. Earlier today there was another problem, where someone accidentally unplugged an entire rack of Apache servers :D (source #wikimedia-tech and the server admin log ) —TheDJ (talkcontribs) 23:01, 26 September 2011 (UTC)
Thanks. Any idea when things will be back to normal... save for random tripping over plugs? LadyofShalott 23:10, 26 September 2011 (UTC)
The final stage of the MW 1.18 rollout is scheduled for Tuesday, October 4, and things will probably settle down within a day or so thereafter. ~ 14:19, 27 September 2011 (UTC)

"edit toolbar" vs "enhanced editing toolbar"... slowness of the enhanced toolbar

There are two editing toolbar that are selectable under the "editing" tab in preferences. It seems like selecting the enhanced toolbar overrides the "edit toolbar". This is a little confusing when both are selected and you unselect the "edit toolbar" and nothing changes.

Also, the "enhanced toolbar" loads unacceptably slow under Firefox 3.6.22. It takes as long as 5 seconds sometime. Any edits made before the page finishes loading, and sometimes you forget to wait, are lost. (Firefox 3.6 is not an obsolete browser so I'm preemptively saying so.) It interferes with editing. This is on a very fast machine running Ubuntu 10.04 LTS fully patched. This problem has existed since Vector was launched. Maybe it's gotten worse but I don't know. I've a"sucked it up" until now but I have to complain about it. Jason Quinn (talk) 23:21, 26 September 2011 (UTC)

Per comments on bug 27620, this slowness is because bug 27488 is still not fixed. You can vote for it!
See also the links provided on mw:Thread:Talk:ResourceLoader/Version_2_Design_Specification/Top-loaded_site_scripts. Helder 00:52, 27 September 2011 (UTC)
There is no need to vote on it. That bug is added to the RL 2.0 project which means it will be taken care of by the time RL 2.0 is ready, which is a high-priority project on itself. The order in which the specification items are done is not really of importance here since it will be deployed as a whole anyway. Right now the RL2.0 developers (including me) are working on the new and improved gadget repositories allowing shared gadgets and a nice visual gadget manager. The next thing is likely going to be the research into top loading and changing the loader system from a "top/bottom" separation (where "top" is a list of blocking loads in the <head> and "bottom" is a list of async/simultaneous loads in the bottom of "<body>"), to a system with "head/other", where "head" is a list of blocking loads in the <head> and "other" queue starts loading asynchronously as soon as the <body> tag opens and is executed right away or on document ready (decided by the module).
The most notable change is that instead of starting to download after the document is ready and then executing it, it will download simultaneously while the body contents are downloaded as well and executed right away as soon as the bottom is reached (instead of starting the download when the bottom of the body is reached). Krinkle (talk) 16:59, 27 September 2011 (UTC)
Wow! Thanks, Helder. After I wrote my comment, I thought that perhaps my comment is too vague for anything to be done about it or for other editors to know what I meant. I regretted posting it. But... your reply was just about exactly what I was looking for without even knowing what answer I hoped to get! I know now that's it's a known issue and where to watch to follow any updates. Thanks a lot! Jason Quinn (talk) 15:25, 27 September 2011 (UTC)

Another hiccup in the system, I think. On the Edit toolbar, there is an icon for Insert. "To a wiki page" and "To an external web page" are at the bottom when you use this icon. Normally, if you copy a Wikipedia URL and paste it there, a secondary box will pop up and ask you if you want an Internal or External link. Indicating Internal will give you a link like this: National Register of Historic Places listings in Shelby County, Tennessee. As of today, if you do that, the box does not pop up asking you what kind of link. So, you pre-select wiki page button at the bottom, and you get this: [Register of Historic Places listings in Shelby County, Tennessee]. Sometimes the copy and paste is preferable to typing in a name, just to insert a link. What's different today in how it works? Maile66 (talk) 15:43, 27 September 2011 (UTC)

Gadgets

Where do I find out who maintains the gadget "Adds two new dropdown boxes below the edit summary box, with some useful default summaries"? It does not input a logical summary when you use a minor edit summary. Thanks--Gilderien Talk|Contribs 19:48, 27 September 2011 (UTC)

Start on Special:Gadgets, look for description above, find MediaWiki:Gadget-defaultsummaries.js, check history. — AlexSm 20:08, 27 September 2011 (UTC)

Is it possible to remove automatic new line after image using template:wide image?

Why?
"
 
Planets and dwarf planets of the Solar System. Sizes are to scale, but relative distances from the Sun are not.
The above image is taken from Wikipedia's entry on the Solar System, and I'll give you five seconds to point out as many flaws as you can. All done? Where do we start? Clearly our sun is dying, its once-dazzling surface now an ember, and there's some other star, several times larger / closer / brighter than our own sun, tucked just out of view in the upper right frame. To be honest, I find that kind of geocentrist shading interesting more than anything else, but it's not what we're here to discuss. Similarly, I'm going to ignore the presence of "dwarf planets", which everyone knows is a concession by International Astronomical Union to keep Arizona happy. No, I'm talking about the fact the planets seem to be breathing down one another's necks, Jupiter within fist-bumping distance of Mars, the asteroid belt apparently slipped from the gaunt hips of our emaciated sun. Now, there's a very good reason that the planets are often presented squished up together like this: it's because black ink is really expensive. Given the choice between illustrating planets as pixel-sized dots on a single page, or going all-in on a 30 page wide fold-out showing planets in all their glory and scale, most artists prefer to cut out all that "empty" space and bring celestial bodies into frame. It's an obvious design solution, but one that nevertheless impacts upon the public's understanding of astronomy. Even though the Wikipedia page makes pains to point out that the scale in this image has been messed about, the industry-wide practice of moving planets about trickles down into public consciousness." From Five iconic science images, and why they're wrong

Template:Wide image is unusable with multiple images (see User:Bulwersator/test2) User:Bulwersator/test is better but it is ugly and results in wiiiiiideeeeee page (pl:User:Bulwersator/test is only ugly). Is it possible to remove automatic new line after image using template:wide image? Bulwersator (talk) 07:24, 28 September 2011 (UTC)

That template is not intended for multiple images. You want something like {{Multiple image}}. ---— Gadget850 (Ed) talk 10:54, 28 September 2011 (UTC)

edits not showing up

Hi,

I am trying to update the page for my organisation which is rather badly out of date.

I've made the updates, and clicked Save, but now I visit the page, twenty minutes later, the old version of it has come up.

How do I get my edits to last?

James Murphy Director of Communications National Youth Orchestra

Page in question: http://en.wikipedia.org/wiki/National_Youth_Orchestra_of_Great_Britain — Preceding unsigned comment added by NationalYouth Orchestra (talkcontribs) 09:48, 28 September 2011 (UTC)

Your addition was automatically reverted.[12] You have other issues and I am taking them to your talk page. ---— Gadget850 (Ed) talk 10:42, 28 September 2011 (UTC)

there is no attribute "class"

The W3C Markup Validation Service is giving this error on every page:

<poem<

there is no attribute "class"

<html lang="en" dir="ltr" class="client-nojs" xmlns="http://www.w3.org/1999/xhtml"...

</poem>

I am guessing this is part of the RTL/LTR updates? ---— Gadget850 (Ed) talk 21:03, 28 September 2011 (UTC)

It looks like our DOCTYPE doesn't allow a class on the html element. The following also generates an error in the validator:
<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Transitional//EN" "http://www.w3.org/TR/xhtml1/DTD/xhtml1-transitional.dtd">
<html class="foo" xmlns="http://www.w3.org/1999/xhtml">
<head>
<title>Title</title>
</head>
<body>
<p>Text</p>
</body>
</html>
(The DOCTYPE is copied from Wikipedia.) Possibly recent changes to MediaWiki introduced a class for the html element. Ucucha (talk) 21:22, 28 September 2011 (UTC)
I didn't think you could have a class= on the <html> - after all, the CSS files are brought in during the <head>...</head> (in our case, with the <link /> elements), which hasn't been parsed yet. --Redrose64 (talk) 21:31, 28 September 2011 (UTC)
I don't see class listed as an attribute for <html>.[13] ---— Gadget850 (Ed) talk 23:43, 28 September 2011 (UTC)
See bug 30497. Reach Out to the Truth 00:34, 29 September 2011 (UTC)
Hi, I'm the developer who created that patch. There's no technical reason why you can't have a class tag on the <html> tag, CSS is applied to all elements, no matter what their position. I wasn't aware that it would cause a failure in validation, so I'm happy to fix it so that it will be applied to <body> instead. Johnduhart (talk) 01:15, 29 September 2011 (UTC)

5,000-revision deletion limit counted incorrectly

On List of Wrestling Observer Newsletter awards, the page said that there were "3,734 deleted edits" while deleted, but when undeleted (or before it was deleted) says that it has over 5,000 revisions and can only be deleted by a steward. What is the reason for this discrepancy? -- King of 22:40, 28 September 2011 (UTC)

Wikipedia:Village pump (technical)/Archive 84#Upper page limit for deletions may answer your question. Jenks24 (talk) 22:49, 28 September 2011 (UTC)

Is this *ever* going to be fixed?

I find it incredible that a massive and popular site like Wikipedia still suffers from this stupid bug whereby stale pages are regularly shown and/or pages are incorrectly shown as uneditable until one adds that stupid "?action=purge" thing to the URL. What a total load of nonsense. There is even a page somewhere I think that explains this "purging" crud as if it was some kind of a beneficial feature rather than a ridiculous bug that should have been fixed years ago.

The usual disclaimers that I know WP is free, built mostly by volunteers, etc., etc., all apply, so there is no need to trot all that out. Just someone PLEASE fix it!!! 109.151.36.98 (talk) 00:30, 25 September 2011 (UTC)

It appears nobody filed a bugzilla request after my above post at #Article appears protected when it isn't. I will probably file one next week when I have more time if it hasn't been done by then, but I don't know how hard it would be to solve the problem. By the way, the purge function is old, has other uses and was not created to fix instances of this problem but is able to do it. PrimeHunter (talk) 00:44, 25 September 2011 (UTC)
(That section link is now WP:Village pump (technical)/Archive 93#Article appears protected when it isn't.) --142.205.241.254 (talk) 22:27, 26 September 2011 (UTC)
I found the old bugzilla:27978 from March 2011 where the problem may have started. The bug was quickly closed as "Resolved" because the problem could not be reproduced on the reported pages. I have reopened the bug with links to many reports since then. PrimeHunter (talk) 22:47, 2 October 2011 (UTC)

Templates not transcluding at Barack Obama

(Please see Talk:Barack Obama#Templates) Transclusion of the final thirteen templates at Barack Obama has failed and, instead, regular links to the templates are being displayed. What could be causing this? Thank you, -- Black Falcon (talk) 17:13, 26 September 2011 (UTC)

Probably Wikipedia:Template limits—the page is transcluding too many and too large templates. Ucucha (talk) 17:21, 26 September 2011 (UTC)
Specifically, it's the post-expand include size:
NewPP limit report
Preprocessor node count: 282439/1000000
Post-expand include size: 2048000/2048000 bytes
Template argument size: 879412/2048000 bytes
Expensive parser function count: 24/500
(Those numbers are in the HTML source of the page.) Ucucha (talk) 17:22, 26 September 2011 (UTC)
Thank you, that clarifies the cause of the problem. I noticed, too, that the article now appears in Category:Pages where template include size is exceeded. Can anything, other than removing a few templates from the article (possible but not ideal), be done to correct this? -- Black Falcon (talk) 23:48, 26 September 2011 (UTC)
It's either that or making the templates smaller. {{Navbox}} and its subtemplates, for example, are huge and could perhaps be streamlined. Ucucha (talk) 23:54, 26 September 2011 (UTC)

I have offered a possible fix at Talk:Barack Obama#Templates), which reduces the template processing counts to:

NewPP limit report
Preprocessor node count: 205464/1000000
Post-expand include size: 1572569/2048000 bytes
Template argument size: 673222/2048000 bytes
Expensive parser function count: 10/500

As this fix may cause problems in future maintenance, I have not implemented it immediately, but described it on Talk:Barack Obama#Templates, so that editors there can decide what they want to do.

--NSH001 (talk) 15:50, 2 October 2011 (UTC)

This gadget no longer works; see #Edit summary max length and behaviour above, and try the gadget out for yourself - it makes no difference.

Could an admin please delete these two pages and remove the gadget's entry from MediaWiki:Gadgets-definition (second entry under "editing" section)? There is no need to fool users into believing they can obtain longer edit summaries when they in fact cannot. — This, that, and the other (talk) 03:51, 27 September 2011 (UTC)

Done. Ucucha (talk) 11:28, 27 September 2011 (UTC)
There is no need to delete MediaWiki:Gadget-LongEditSummaries and MediaWiki:Gadget-LongEditSummaries.js. Can you please undelete those so that the code can still be seen by others ? The existance of those pages alone doesn't have any influence on the Gadgets system. Assuming you want someone to be able to fix/redo that script... Krinkle (talk) 17:06, 27 September 2011 (UTC)
I've restored. Good luck fixing the script. :) Is there anything that needs to be done to the page to indicate that it is being retained for this reference only and that it doesn't work? </clueless> --Moonriddengirl (talk) 15:40, 3 October 2011 (UTC)

Template Question, time since last edit

Is it possible to have a template add a page to a category based on how long it has been since the page was edited? For example, if the template is placed on a page at the time of page creation, would it be technically possible for it to add the page to a category 30 or 90 days after the last edit to the page by anyone? If so, will the category addition occur even if no one ever again accesses the page, or would the cached version, with no category, remain until the first time it was accessed after the timer ran out? Monty845 23:56, 27 September 2011 (UTC)

Yes, I think this could be done using two of the magic words, {{REVISIONTIMESTAMP}} and {{CURRENTTIMESTAMP}}. But the software would not realise that the page needed to be re-built and re-categorised after N days, so the idea wouldn't work especially well. Also, depending on what you were hoping to use the template for, the scheme might be broken by an automated/semi-automated "gnome" edit that didn't affect the part of the page you had in mind. -- John of Reading (talk) 11:54, 28 September 2011 (UTC)
Thanks for the reply, the lack of re automated rebuilding would stop the idea from working, which is what I feared. Monty845 18:59, 28 September 2011 (UTC)
You could employ a bot to make null edits on those pages every day, I know we have such a bot in my home wiki. — AlexSm 20:03, 28 September 2011 (UTC)
This kind of bot has been proposed before and rejected. There is a reason the software works in the way it does, using null-edits to force updates is not the correct way to go about things. - Kingpin13 (talk) 20:20, 28 September 2011 (UTC)
You shouldn't need null edits for this - I recently added this exact feature to create Category:Stale Userspace drafts - and the category filled up with 12k pages in no time. Avicennasis @ 14:34, 1 Tishrei 5772 / 14:34, 29 September 2011 (UTC)
Your edit to {{Userspace draft}} made the software re-cache all the pages using that template. Going forwards I think you will have to make a monthly edit to the template to make the software rethink the categories for each draft. -- John of Reading (talk) 15:39, 29 September 2011 (UTC)

What is the current standard for adding little status symbols for iw links? FA and GA symbols are added, how about the "decent article" status that's used in da, fi and sv Wikipedias. The status equals roughly the B quality in here, being a status with more relaxed standards and more casual promotion process than GA or FA. Pitke (talk) 14:11, 28 September 2011 (UTC)

Just worked this out. You are referring to the {{link GA}} and {{link FA}} templates which put   and   symbols on the left of the "languages" list (or, if in MonoBook skin, the templates which amend the style of the bullets in the same list), as seen against "Deutsch" and "Español", etc. on Vietnam war. This seems to be done via css; the effect of {{link FA|es}} is to add an empty HTML element: <span id="interwiki-es-fa"></span> so I personally can't do anything to add the equivalent for this class (which seems to be exclusive to the Danish, Finnish and Swedish Wikipedias). --Redrose64 (talk) 19:36, 29 September 2011 (UTC)
Ok, thanks. All that remains is to probe whether such symbols would be welcome in a wiki that doesn't use the classification itself. Kinda. Pitke (talk) 10:39, 30 September 2011 (UTC)
Have found in MediaWiki:Common.js that there is function LinkFA() - this seems to be highly relevant here. At the top of that function there is a note that it's maintained by R. Koot (talk · contribs), so you could try contacting him. --Redrose64 (talk) 23:33, 1 October 2011 (UTC)

large numbers are rendered differently by various servers, leading to number formatting errors.

Large numbers, such as 82000000 are getting rendered as 8.2E+7 by some server (srv*) and as 82000000 by others (wm*). The 8.2E+7 notation causes templates that format numbers (such as {{val}}) to fail and pages using these templates to be rendered incorrectly. It seems obvious that all server should return the same results in order to be able to create reliable rendering of large numbers. Assuming there is no use case for 8.2E+7 notation, I'd like to get all servers to return 82000000 when you type 82000000.

To test this, you can install this Chrome extension I created: it takes the name of the server that rendered each Wikipedia page from the comments inside the HTML and makes it visible at the bottom of the page. Alternatively, you can use view-source to see this information manually. Then try refreshing this page to see how the below test is rendered by various servers:

{{#expr:82000000}} = 82000000

Any help in tracking down people that can help resolve this is appreciated.     SkyLined (talk) 21:55, 28 September 2011 (UTC)

NB: I just found that not all srv* server render it as 8.2E+7: some do, some don't.    SkyLined (talk) 21:57, 28 September 2011 (UTC)
This is a known issue, I believe; some of the servers are configured differently than others, and this causes some math operations to return different values depending on the server. Ucucha (talk) 00:43, 29 September 2011 (UTC)
This is, AFAIK, ultimately down to features of the operating system and hardware, since MediaWiki uses PHP maths functions, PHP uses the system's native C functions, and the C functions use direct OS/assembler-level maths. Some of the modern servers are 64-bit architecture while the older ones are 32-bit, which probably makes a substantial difference. Essentially not something which is likely to be resolved. Happymelon 20:57, 29 September 2011 (UTC)
A long time ago (three years ago, or some such), I recall being told that many changes like this would go away once all the servers were updated to run the same version of PHP, which apparently wasn't the case at the time. Is there any way to tell which version(s) of the infrastructure software are being run currently? Dragons flight (talk) 21:30, 29 September 2011 (UTC)
Yes: Special:Version section "Installed software". --Rogerhc (talk) 23:15, 2 October 2011 (UTC)
That only tells you the software on the machine that happens to render it. I was looking for a way to see software usage across the entire production cluster. As established in the bugzilla thread, some machines are using PHP 5.2.4 and others are using 5.3.2. Dragons flight (talk) 23:34, 2 October 2011 (UTC)
If you want faster/consistent servers, visit foundation:Fundraising. --Redrose64 (talk) 21:27, 29 September 2011 (UTC)
Helpful, Redrose, except for the complete and utter absence of any analysis of a connection between a lack of funding for servers and the issue under discussion. Is there a bugzilla bug filed on this subject, does anyone know? "Not something which is likely to be resolved" somewhat lacks the can-do spirit --Tagishsimon (talk) 21:31, 29 September 2011 (UTC)
Funny, that. They are more likely just hire more HR people. Or perhaps another storyteller! Killiondude (talk) 23:01, 1 October 2011 (UTC)
My half-informed guess is that this is likely to be caused by the lucid upgrade. This upgrade is being rolled out to all servers in a staggered fashion, so if we don't fix it, this problem will only get worse over time, as the probability that you'll hit a server that wrongly formats 82000000 as 8.2E+7 rises and eventually becomes 100%. I tried to ping the guy who's doing the lucid upgrades so he can take a look, but since he's not around and since I should really go to sleep now, it's probably best to file this as a bug in Bugzilla. We like bug reports for tracking purposes anyways. --Catrope (talk) 23:14, 29 September 2011 (UTC)
I'm happy to file a bug if you'll tell me where to find the Bugzilla.     SkyLined (talk) 23:19, 29 September 2011 (UTC)
Nm; found it using google at [14]     SkyLined (talk) 23:26, 29 September 2011 (UTC)
Filed bug 31259.     SkyLined (talk) 23:37, 29 September 2011 (UTC)

Bing isn't updating urls

I moved Bunker (golf) to Hazard (golf) on 14 May 2009. Although Bing has updated the article title in its search results, clicking the link still takes you to the redirect. How do we get Microsoft's attention on this? Marcus Qwertyus 08:42, 29 September 2011 (UTC)

Creating new usernames

A current WP:ANI discussion about a longterm anticontributor leads me to wonder: how does one create an account? I thought that there were only two ways to create accounts (either by creating the account with itself, as I did, or by becoming an AccountCreator), but he's obviously figured out some other way to do it, since surely we didn't give a longtime vandal the AccountCreator right. Nyttend (talk) 04:58, 30 September 2011 (UTC)

If I'm right, the Account Creator flag simply removes a limit of 6 accounts/IP address/day. Any user that isn't blocked with "account creation blocked" enabled can create accounts, but anyone who wants to make more than 6/IP address/day needs to have the flag.Jasper Deng (talk) 05:01, 30 September 2011 (UTC)
Indeed. Any account which isn't ACB blocked can register more just by visiting signup whilst logged in. -- zzuuzz (talk) 06:09, 30 September 2011 (UTC)
It's that signup page I've never before seen. Is it linked from somewhere? I can't get useful results from WhatLinksHere. Nyttend (talk) 11:24, 30 September 2011 (UTC)
Special:UserLogin/signup also. The signup is just a version of, and linked from, Special:UserLogin. It's linked specifically from some username block templates, and more generally Special:SpecialPages. However all you need to do is have a browser tab open where you aren't logged in and you can copy-paste or click the link into a new tab all you like (up to the account limit). I expect this vandal also has it has a bookmark. -- zzuuzz (talk) 11:38, 30 September 2011 (UTC)
The main problem with trying to use Special:WhatLinksHere in this manner is that it it doesn't work for virtual namespaces - those with a negative number, such as Special: pages. Another problem is that it only lists links established using wikilinks, not the http: form. --Redrose64 (talk) 14:44, 30 September 2011 (UTC)
The first problem is bug 17597. You can vote for it. Helder 17:15, 30 September 2011 (UTC)

Problem with Template:Infobox election

Someone in there, there seems to be a problem with a colspan being one greater than it should be. Specifically, if you render United Kingdom general election, 2005 in PDF, the map seems to span one column more than candidates.

I've been trying to figure out where this extra column is comning from for the last few hours, to no avail. Can someone confirm/deny if the colspan is set to 4 throughout the infobox? Headbomb {talk / contribs / physics / books} 06:33, 30 September 2011 (UTC)

There is no column span problem, and no extra column. It appears that the PDF renderer ignores the CSS width specified on the tables. So while the browser displays the nested tables at 100% of the width of the outer table, the PDF version sizes it to fit its content and (since the cell content is left-aligned) it leaves empty space on the right. This sounds like something to bring up on Help:Books/Feedback. Anomie 11:49, 30 September 2011 (UTC)
Ah nested tables. Should have known. Headbomb {talk / contribs / physics / books} 15:39, 30 September 2011 (UTC)

dual wikipedia pages

I am in a discussion about the aspartame controversy at NPV. If I do not sign on the discussion is not visible. It appears when I sign on. Is this normal? Do you have dual pages? — Preceding unsigned comment added by Arydberg (talkcontribs) 07:16, 30 September 2011 (UTC)

No, we don't have dual pages, but sometimes an old version gets cached incorrectly. I have purged the page so that the latest version will be copied again to all the Wikipedia servers. You could also try bypassing your browser cache. -- John of Reading (talk) 08:02, 30 September 2011 (UTC)
Logged in users are accessing pages using a more frequently updated server while IPs use a server that relies more heavily on cached content. Killiondude (talk) 22:50, 1 October 2011 (UTC)

DNS entry hacked?

Did the main DNS entry get hacked? It whois now shows it as

Domain names in the .com and .net domains can now be registered
with many different competing registrars. Go to http://www.internic.net
for detailed information.
WIKIPEDIA.ORG.ZZZZZ.GET.LAID.AT.WWW.SWINGINGCOMMUNITY.COM
WIKIPEDIA.ORG.IS.WRONG.GO.4.GULLI.COM
WIKIPEDIA.ORG

And you can't connect. Vegaswikian (talk) 18:57, 30 September 2011 (UTC)

It came back for a few and then went down again. Vegaswikian (talk) 19:08, 30 September 2011 (UTC)
Where do you get that? whois.org and command-line whois wikipedia.org show normal results for me, and connecting to the site works fine. Ucucha (talk) 19:23, 30 September 2011 (UTC)
That's known as "whois spam", and has nothing to do with DNS. Some methods of querying the whois database search for any record containing the entered text (rather than just an exact match), so some spammers include the names of high-profile websites in their own records in some strange attempt to get traffic to their sites. Anomie 19:25, 30 September 2011 (UTC)
I got that from SamSpade. http://www.internic.net/whois.html also seems to be having trouble finding the site with their whois function. But still no access. Vegaswikian (talk) 19:28, 30 September 2011 (UTC)
For a period commencing approx. 19:00 UTC I was unable to access any Wikipedia page - the error generated by Firefox was something like "Firefox is unable to locate the server at en.wikipedia.org" and the usual "make sure it's spelled correctly" stuff. --Redrose64 (talk) 19:46, 30 September 2011 (UTC)

Bandwidth throttling

Does wikipedia use bandwidth throttling, and if yes which conditions engage/disengage it ? Gzilirion (talk) 07:00, 1 October 2011 (UTC)

To my knowledge, there is no throttling of any kind. Edokter (talk) — 11:20, 1 October 2011 (UTC)
Wikipedia:Database_download#Sample_blocked_crawler_email Bulwersator (talk) 11:23, 1 October 2011 (UTC)

Suggestion re some images

Would it be feasible to have a prefference option so that stuff tagged as {{badimage}} will not display, thumbnail or make http requests for the image at all?

This would be of great assistance to certain users that don't want to see that sort of content. Sfan00 IMG (talk) 10:52, 1 October 2011 (UTC)

https://bugzilla.wikimedia.org/show_bug.cgi?id=31298 - I've put in an 'enhancement' request.. Sfan00 IMG (talk) 19:04, 1 October 2011 (UTC)

Auto-include talk header template

I suggest that the wiki software auto-included {{talk header}} at the top of every talk page at the moment of its creation. It's a really good reminder for experienced editors and a good intro for new editors. Maybe even create a bot to add it at the top of every page that lacks it now. Thanks for listening! Woz2 (talk) 17:10, 1 October 2011 (UTC)

If we want that thing on every page, let's do it in some sensible way—sitewide JavaScript or a MediaWiki message—not by transcluding a template on every single talk page on this project. Ucucha (talk) 17:13, 1 October 2011 (UTC)
The documentation currently states "This template should be used only when needed. There is no need to add this template to every talk page." Has consensus changed on that? I hope not. Anomie 17:31, 1 October 2011 (UTC)
It also states that it is "intended to be used on particularly active talk pages that attract commentary from inexperienced editors, and/or high levels of debate from everyone", none of which can possibly apply to non-existent pages. --Redrose64 (talk) 18:18, 1 October 2011 (UTC)

Hmmm... What's the downside of including it? Woz2 (talk) 17:39, 1 October 2011 (UTC)

Clutter. At one time, absent talk pages were commoner than talk pages that existed; so the "discussion" link being coloured blue could be seen as an indicator that discussion had started. Then WikiProject banners were invented, and so the red "discussion" link is a rarity. Many talk pages (like this one) consist of nothing but project banners, there is no discussion as such. These are tolerated, but I doubt that a general use of {{talk header}} would be. --Redrose64 (talk) 18:18, 1 October 2011 (UTC)

Ok Thanks! Woz2 (talk) 18:23, 1 October 2011 (UTC)

Server error

I keep getting the following error message when I attempt a contributions search in the Wikipedia Talk namespace:

Proxy Error

The proxy server received an invalid response from an upstream server. The proxy server could not handle the request GET /wikipedia/en/w/index.php.

Reason: Error reading from remote server

Apache/2.2.8 (Ubuntu) mod_fastcgi/2.4.6 PHP/5.2.4-2ubuntu5.12wm1 with Suhosin-Patch mod_ssl/2.2.8 OpenSSL/0.9.8g Server at secure.wikimedia.org Port 443

Does anyone have any idea why this is happening? ---RepublicanJacobiteTheFortyFive 18:12, 1 October 2011 (UTC)

I am still waiting and hoping for an answer to this question, because I am still getting this error message when I attempt to do a contributions search in the Wikipedia Talk namespace. I have not been able to conduct such a search in days, and it is rather frustrating. ---RepublicanJacobiteTheFortyFive 16:30, 2 October 2011 (UTC)
Try using the new HTTPS servers NEW server. —TheDJ (talkcontribs) 17:02, 2 October 2011 (UTC)
Yeah, that connected right away. Thanks! ---RepublicanJacobiteTheFortyFive 17:38, 2 October 2011 (UTC)

Edit notice on BLP articles

  Resolved
 – With many thanks for the friendly help from David Göthberg, TheDJ, Anomie, Edokter, and Redrose64. Now all BLP edits show the important notice, rather than just a minority of them. First Light (talk) 15:04, 2 October 2011 (UTC)

There is a very helpful edit notice that was created for Biographies of Living Persons, but unfortunately it only shows up when you click the "edit this page" at the top of the article. Never when you click to "edit" a section of a BLP article. Go to Fareed Zakaria, the most recent BLP I edited, to see what I mean. All BLP articles exhibit this bug. Since the vast majority of edits are made to individual sections, can this be fixed so that the edit notice appears on all edits made to BLP articles? Thanks, First Light (talk) 19:43, 1 October 2011 (UTC)

This is a known problem, and it's been mentioned before. The relevant JavaScript is MediaWiki:Common.js and the notice is Template:BLP editintro; the doc page for that states "this edit intro is shown automatically when editing a page categorized as either Category:Living people or Category:Possibly living people, provided that the edit page is accessed through the main edit tab". See also WP:EDITINTRO and MediaWiki talk:Common.js/Archive 16#BLP editintro. --Redrose64 (talk) 20:08, 1 October 2011 (UTC)
I see. Too bad, since it does very little good the way it stands. If this could be fixed, it would help with one of Wikipedia's worst problems (certainly the worst public relations problem): people using Wikipedia articles to attack their enemies. See an example here.[15] First Light (talk) 20:43, 1 October 2011 (UTC)
If somebody is determined to defame another, they will ignore the {{BLP editintro}} whether it's shoved in their face or not. Wikipedia:Vandalism gives advice on spotting and dealing with cases like this. --Redrose64 (talk) 21:27, 1 October 2011 (UTC)
But for new editors and vandals, it will make them think twice, or maybe find a reference. I've also seen both new and long-time editors unknowingly violate BLP policies. As far as the PR issue, it only confirms to the press that we aren't doing everything we can to protect the reputations of innocent people. Though I do very much appreciate what we are doing on Wikipedia with limited volunteer technical time and expertise. First Light (talk) 21:47, 1 October 2011 (UTC)
Well, I see two ways we can fix this:
1: We could use a for-loop: Currently the code in MediaWiki:Common.js looks for Category:Living people and Category:Possibly living people and if they exist then adds "&editintro=Template:BLP_editintro" to the edit link at the top of the page. This happens when we view the page, before we click the edit link. As far as I can see we can add a for-loop that locates all the section edit links since they are clearly marked with <span class="editsection">, and then add the same "&editintro=Template:BLP_editintro" to all those section edit links.
2: We could use a hidden category and an ajax call: I have noticed that hidden categories are visible on the edit page even if we edit another section than the one that adds the category. So we could add a hidden BLP category to all BLP pages (perhaps automatically by using the infoboxes). Then we can use javascript that runs when we edit the page (not before we edit the page) and checks for that category and then uses an ajax call to render {{BLP editintro}} and inserts it.
I prefer method one above (for-loop) since it is simpler to implement, it costs less server resources, and I am sure it will work. While I don't know enough about ajax coding to even be sure that method two would work. However, my javascript skills are too rusty and I am semi-retired from Wikipedia so you guys need to ask the experts over at MediaWiki talk:Common.js to implement it.
--David Göthberg (talk) 23:21, 1 October 2011 (UTC)
Thank You! A possible solution for something that is worth solving. I'll put a note over there, including a paste of your message, to see what they can do. Regards, First Light (talk) 23:39, 1 October 2011 (UTC)
Done. —TheDJ (talkcontribs) 12:37, 2 October 2011 (UTC)
Well done, and thanks. First Light (talk) 14:57, 2 October 2011 (UTC)

Dis needed in Template:WikiProject Royalty

Hmmm, how do I disambiguate royalty to royal family in Template:WikiProject Royalty? I am not seeing it linked there... --Piotr Konieczny aka Prokonsul Piotrus| talk to me 01:51, 2 October 2011 (UTC)

It uses {{WPBannerMeta}}, which automatically generates the link (and all the rest of the box). In this case, you'd want to use |MAIN_ARTICLE=[[Royal family|Royalty]] (if there is consensus for the change, of course). Anomie 02:00, 2 October 2011 (UTC)
Similar to this edit which I did for WikiProject Mills over a year ago. --Redrose64 (talk) 12:47, 2 October 2011 (UTC)

Show recent edits to pages I've edited

I'd like a page that can do something like "show me all recent edits to pages I've edited in the last n days", where "recent" is defined as "since the last time I edited that page". For bonus points, it could exclude pages where the only edits I made were marked as minor. Does anyone know of such a thing? Toohool (talk) 05:41, 2 October 2011 (UTC)

Use " Add pages I edit to my watchlist"? Bulwersator (talk) 05:52, 2 October 2011 (UTC)
The problem with that is that I don't want to watch those pages forever. If I leave a message on someone's talkpage, for example, I don't want to see every message that someone else leaves until forever. Toohool (talk) 06:03, 2 October 2011 (UTC)
See Wikipedia:Village pump (proposals)/Archive 77#Functionality on "my contributions" for a discussion of User:Markhurd/hidetopcontrib.js, which approaches the functionality you want.-gadfium 07:49, 2 October 2011 (UTC)
That's somewhat helpful, but doesn't fully solve the problem. After I see that I'm no longer the top edit and look at the history of a page, there's no further indication if another edit happens. And later on, I might forget that I already looked at that page. The ideal tool I'm looking for would show all edits subsequent to mine, sorted by most recent. Toohool (talk) 17:44, 2 October 2011 (UTC)
History pages over at Commons show a marker reading updated since my last visit against recent edits. --Redrose64 (talk) 18:11, 2 October 2011 (UTC)
Sounds like the sort of thing that could be done using the API and a a script in your .js page. Basic idea would to create a dynamic "watchlist" from pages you recently edited and use the API to check for changes. Changes could be shown in a separate section of your watchlist.
I'd be interested in trying it. I don't think it would be too hard. It would be useful. I'm busy for the next couple of days, however. --RA (talk) 19:03, 2 October 2011 (UTC)

Updating Template:Tel Aviv suburban railway map

Recently, a new line was openned by Israel Railways. Unfortunately, I doin't quite undrstand how to edit {{Tel Aviv suburban railway map}} correctly. Please update it in the following way:

  1. The line from HaRishonim ends at TA Savidor Central.
  2. The line from Hod Hasharon Sokolov goes to the end of Tel Aviv (Tel Aviv HaHagana), and then goes to the following stations (all new, not accessable):
    1. Holon Junction Railway Station
    2. Holon-Wolfson Railway Station
    3. Bat Yam-Yoseftal Railway Station
    4. Bat Yam-Komemiyut Railway Station
    5. Rishon LeZion Moshe Dayan Railway Station

An updated map on the company's site can be found at http://rail.co.il/HE/Stations/Map/Pages/RouteMap.aspx. עוד מישהו Od Mishehu 15:02, 2 October 2011 (UTC)

  Doing... I'll take this - I've amended many RDTs in the past. --Redrose64 (talk) 15:27, 2 October 2011 (UTC)
  Done Five new stations, but only three new rows required. --Redrose64 (talk) 16:19, 2 October 2011 (UTC)

I can't access mobile.wikipedia.org since October 1, 2011. My browser gives me a parser error.

I am using a Samsung SCH-U430 cell phone to access the moblle version of Wikipedia. The URL that I have always used is http://mobile.wikipedia.org. All of a sudden the link won't work today, I get the error message "Parser Error". My phone is using Obigo Browser Q04C1-1.22 built on Apr 27 2010. I have tried other URL's such as m.wikipedia.org, en.mobile.wikipedia.org, and en.m.wikipedia.org. They still give me the same parser error. I also notice that that mobile Wikipedia site is a beta version. Can anybody help me out to get mobile Wikipedia working again on my phone? All my other bookmarks and other websites that I go to work as before. — Preceding unsigned comment added by 70.81.58.142 (talk) 16:34, 2 October 2011 (UTC)

This is because the old WAP 1.0 server (mobile.wikipedia.org) was switching to the new m.wikipedia.org mobileFrontend. This mobile frontend does support wap, but it only paginates per header, not by a fixed size. So likely the server is sending way too much data towards the user, and probably invalid WML as well, causing these errors. I think this switch should be undone, the new fronted is totally not ready to support WAP yet, as I have told the team multiple times. —TheDJ (talkcontribs) 17:07, 2 October 2011 (UTC)
bugzilla:31310TheDJ (talkcontribs) 20:27, 2 October 2011 (UTC)

Until this issue is fixed I've been using http://wapedia.mobi/en/ — Preceding unsigned comment added by Barrymtl (talkcontribs) 17:26, 3 October 2011 (UTC)

Mobile fontend general (maybe dumb) question

I'm posting this here 'cause you guys are quickest to respond: How does the content for the mobile frontpage end up at http://en.mobile.wikipedia.org/ ? The one @nv is empty right now, and I'm trying figure out where I'd need to put whatever should be shown there. I found the headers in the system-messages, but where's the body of the blurp/how does it get there? Choyoołʼįįhí:Seb az86556 > haneʼ 20:41, 2 October 2011 (UTC)

One way to have a mobile main page is to add ids beginning with "mf-" to elements that should be displayed on mobile (ie <div id="mf-foo">content</div>). (This isn't the method used by enwp.) See m:Mobile Projects/Mobile Gateway. --Yair rand (talk) 22:52, 2 October 2011 (UTC)
(e/c) http://meta.wikimedia.org/wiki/Mobile_Projects/Mobile_Gateway#Selectors should help you. I don't think you need to add a bugzilla ticket anymore, just add the selectors to the relevant sections. You can look at the (normal) english mainpage with the "View source" option of your browser to see how it was done for en.wp or use the mf- method. —TheDJ (talkcontribs) 22:57, 2 October 2011 (UTC)
AH! good pointer. thanks. Choyoołʼįįhí:Seb az86556 > haneʼ 23:38, 2 October 2011 (UTC)

PDF generation does not work

See WP:HD#Render Server Error since October 2, 2011. It seems articles can no longer be downloaded as PDF. Does anybody know what the problem is and when it will be fixed? Toshio Yamaguchi (talk) 17:46, 3 October 2011 (UTC)

watchlist tweak: is this possible?

It seems to me it should be possible to tweak a watchlist so that pages I have been active on in the last X days are highlighted (or possibly even separated into their own section). That would save a lot of digging through stuff I'm merely watching to get to stuff where I'm currently an active participant. There's no way to do this through CSS alone, obviously, but does anyone know if there's a javascript tool that can do it? or would this require developer intervention? --Ludwigs2 17:54, 3 October 2011 (UTC)

Personally I've found that using 'my contributions' with the 'hide top contribs' script makes it a lot easier to find my contributions to pages that may need following up on. –xenotalk 17:57, 3 October 2011 (UTC)
Link to Hide Top Contribs. Also, a shameless plug for my own Mark Edits After My Own, which does something similar but instead of hiding Top edits, instead just highlights all pages that are NOT Top and are NOT on your watchlist (therefore drawing your attention to pages that require your attention since you no longer have the most recent edit on that page and the page is not on your watchlist). So I use my contribution page and my watchlist in combination, in this manner, by watching pages with few edits and by not watching pages with plenty of edits and instead using my contributions page for those instead, so that my watchlist doesn't get filled up. Gary King (talk · scripts) 19:53, 3 October 2011 (UTC)

revisiting usbk/usbktop widths

Please discuss my suggestion. ⇔ ChristTrekker 18:32, 3 October 2011 (UTC)

Native HTTPS support enabled

Please help in updating out scripts to no longer use secure.wikimedia.org work-arounds and use protocol-relative urls to Wikimedia domains (i.e. //upload.wikimedia.org instead of http://upload.wikimedia.org.

Thanks, Krinkle (talk) 20:36, 3 October 2011 (UTC)

This might be useful to work from, especially if someone finds a way to filter it by namespace (only template is really interesting). Ucucha (talk) 22:00, 3 October 2011 (UTC)
Couldn't you have just got a single certificate that covered all the wikimedia domains rather than giving to give every domain it's own IP? Plugwash (talk) 04:01, 4 October 2011 (UTC)
No. Certificates don't work that way.--Ryan lane (talk) 18:27, 4 October 2011 (UTC)
Actually, the Subject Alternative Name field can allow a certificate to work that way. Its main limitation, according to a DigiCert web page, is that mobile devices running Palm OS or old versions of Symbian OS do not recognize it. PleaseStand (talk) 20:53, 4 October 2011 (UTC)
Also one thing that doesn't seem to be made clear about the new https system. Is login information protected in any way as it passes between "esams" and "pmtpa"? Plugwash (talk) 04:17, 4 October 2011 (UTC)
The information is not. For the majority of the requests there is no traffic between esams and pmtpa, so any tracking that could happen at that level is pretty limited. We are looking at ways to protect this information as well, though.--Ryan lane (talk) 18:28, 4 October 2011 (UTC)
It looks like the current state on this is that traffic between the Amsterdam proxies and the US-based servers is still unencrypted HTTP (same as they were before, with or without the secure.wikimedia.org gateway). This does in principle leave that traffic vulnerable to sniffing from the US or Dutch governments or a couple of large corporations. This could be changed to run over an encrypted VPN or something, but we've so far considered that a lower priority than protecting against the "hostile local network or internet provider" case -- which SSL on the front-end protects you against. --brion (talk) 18:18, 4 October 2011 (UTC)

Scripts that need updating

These scripts need updating to use protocol-relative URLs...

Hope that helps. —Tom Morris (talk) 13:38, 4 October 2011 (UTC)

I've done as much as i can do. Someone please use the scripts extensively to make sure i didn't break anything. I note that most of these issues would not have existed if people had been using wgScript, wgScriptPath or wgArticlePath. Please people, only use absolute paths in scripts, if you need to go cross wiki. if you stay within the same wiki, we have these perfectly usable global variables, that are there to prevent problems like this and also make you script much more portable for other wiki users. —TheDJ (talkcontribs) 19:40, 4 October 2011 (UTC)
I've added one that doesn't work. →Στc. 04:27, 5 October 2011 (UTC)

JavaScript help needed

I need a snippet of JavaScript that will triggers the p-cactions menu in Vector to show. It must do so by triggering an event like :hover, and must not change any CSS propereties. (I need this to fix a small bug in my MenuTabsToggle gadget, where IE immediately hides the menu after toggling to menus, despite the mouse hovering over the arrow that normally triggers it to show.) Edokter (talk) — 07:49, 4 October 2011 (UTC)

Never mind; I restructured some HTML and it seems the IE bug magically disappeared. Edokter (talk) — 15:03, 4 October 2011 (UTC)

Template issue with Template:Lz12 and Template:Lz

  Resolved
 – The Lz12 template was modified to remove the deleted Lz template. --Kumioko (talk) 15:09, 4 October 2011 (UTC)

The template {{Lz12}} contains a deleted template of {{lz}} and I am not sure what this template affects so before I start messing with it and break something else I wanted to drop it here for comments. --Kumioko (talk) 13:06, 4 October 2011 (UTC)

The original for that deleted template may be found at meta:Template:Lz. --Redrose64 (talk) 13:40, 4 October 2011 (UTC)
Thanks, do you have any advice for how to fix this template? --Kumioko (talk) 13:48, 4 October 2011 (UTC)
{{lz}} only occurs within the "noinclude" part of the template. Changing those lines isn't going to damage anything. I suggest you just remove the three lines from "Compare..." to "123}}". -- John of Reading (talk) 13:59, 4 October 2011 (UTC)
Thank you its done and I have marked this discussion as resolved. --Kumioko (talk) 15:09, 4 October 2011 (UTC)

Talk pages: new users - unable to create page

Unable to create talk page for new users, for example to place CSD and/PROD warnings. --Kudpung กุดผึ้ง (talk) 03:43, 5 October 2011 (UTC)

If you tried with Twinkle, I reported a similar problem above. However, I can still create pages manually. In fact, I just did so. Regards, Orange Suede Sofa (talk) 03:53, 5 October 2011 (UTC)
I believe we just fixed this problem about 10 min ago. Please try it again. Thanks! -- RobLa-WMF (talk) 04:30, 5 October 2011 (UTC)

Twinkle's AIV tab doesn't work either. →Στc. 04:28, 5 October 2011 (UTC)

Enhanced recent changes doesn't work

  Resolved
 – fixed

Enhanced recent changes is no longer working on my account. Is this the case for others, or am I alone in this problem?

I have used the feature for years, without problem. But this morning, it simply does not work. My watchlist appears, with all of the diffs displayed, rather than collapsed. as usual. Other Java features seem to work normally, so I don't think I have a Java problem. Has there been any change made overnight which could have caused this?

I have checked my preferences, and "enable enhanced recent changes" is checked. I have rebooted my computer, in case something hadn't loaded properly; but this makes no difference. Please help, as it is difficult navigating a long watchlist with all diffs displayed. RolandR (talk) 08:12, 5 October 2011 (UTC)

That would be bugzilla:31358, also strange question, What skin are you running?. Peachey88 (T · C) 08:44, 5 October 2011 (UTC)
Vector. Looking at the contributions above, it is clear that others are experiencing this issue, and that it is related to the latest upgrade. I hope it can be resolved soon. RolandR (talk) 08:51, 5 October 2011 (UTC)
It is Vector. Thanks for your help, I'll change my skin for a while. Pitke (talk) 15:20, 5 October 2011 (UTC)
It's working again. But I still have some of the other problems noted below. RolandR (talk) 23:02, 5 October 2011 (UTC)

Missing text when saving

  Resolved
 – markup typo found

Hello, I have just created and saved a draft user page (FionaSturgeon/ChangeBASE), however the saved version does not show all the content I put in at the creation stage. When I go to edit, the text in question reappears, so it definitely is there in some form but is not being displayed for whatever reason. I am using Google Chrome.

Does anyone else have this issue, or know how I go about fixing it? I am new to Wikipedia so it is likely that I'm missed something - any help would be appreciated.

Many thanks, Fiona — Preceding unsigned comment added by FionaSturgeon (talkcontribs) 13:07, 5 October 2011 (UTC)

You are closing all your references with <ref>; you need to start with <ref> and close with </ref>. ---— Gadget850 (Ed) talk 13:21, 5 October 2011 (UTC)

WP:REFNEST has stopped working properly

  Resolved
 – duplicate of #bug in #tag:ref parser function

Help - the code behind WP:REFNEST has stopped working properly, resulting in references embedded within notes displaying a series of random characters. Any ideas? Shem (talk) 14:49, 5 October 2011 (UTC)

See #bug in #tag:ref parser function above. ---— Gadget850 (Ed) talk 14:52, 5 October 2011 (UTC)

Thanks - missed that. Shem (talk) 15:30, 5 October 2011 (UTC)

Twinkle problems

I noticed that Twinkle isn't loading. I cleaned up my js file, but that didn't work. Removing TW from my js file and enabling it as a gadget. I usually use the secure server, so I switched to the regular server, but had no luck with that. At all the times I cleared or bypassed my cache. I run an unbranded version of Firefox 7. Samuell Lift me up or put me down 16:14, 5 October 2011 (UTC)

Many things are not working right now; check the massive section above to see if problems have already been reported. This one has, at #Some_JS_tools_broken. Aren't not-quite-thoroughly-tested new versions of MediaWiki great?! –Drilnoth (T/C) 16:11, 5 October 2011 (UTC)
Interesting right ? Even after having been in production on 3 wiki's for over 3 days, no one but en.wp has found many of these issues. As a user you underestimate the thousands of usage patterns that there are. Testing will only reveal so much. Also, Twinkle is a local tool. The local community is responsible for keeping it maintained. Thus it is no more than expected basically that many of the gadgets break when a release is being made. Dust will settle. —TheDJ (talkcontribs) 16:17, 5 October 2011 (UTC)
See WT:TW#No Twinkle at all? and other discussions on that page. Regards SoWhy 16:20, 5 October 2011 (UTC)
The funny thing is, aside from a few very minor niggling issues that are more cosmetic than anything else, my Twinkle has (so far) been working just fine... - The Bushranger One ping only 19:39, 5 October 2011 (UTC)

Redirects to sections

  Resolved
 – bogus

Is this broken or is it just me? Links like WP:UNDUE and WP:AUTOCONFIRM lead to the correct pages, but are not navigating to the sections. Using Firefox 7.0.1. — Bility (talk) 16:38, 5 October 2011 (UTC)

  Works for me ΔT The only constant 16:39, 5 October 2011 (UTC)
(edit conflict) Nevermind, it was me. — Bility (talk) 16:40, 5 October 2011 (UTC)

Banging my head

MediaWiki:Broken-file-category was introduced with 1.18 and Im trying to configure it so that it only lists articles/templates with missing files. (other namespaces really dont matter too much). But I cannot see to get it configured correctly. ΔT The only constant 17:28, 5 October 2011 (UTC)

I used {{namespace detect showall}} in {{broken ref}} to do similar for Cite errors. ---— Gadget850 (Ed) talk 18:17, 5 October 2011 (UTC)
Can you go ahead and adjust the page accordingly? ΔT The only constant 20:31, 5 October 2011 (UTC)
  Resolved
 – duplicate of #bug in #tag:ref parser function

In Love the Way You Lie#Notes, I have a footnote, which is referenced. However, the link to the reference now shows ?UNIQb4e22634542f5cb-nowiki-00000079-QINU?18?UNIQb4e22634542f5cb-nowiki-0000007A-QINU? instead of the [citation number]. I use Google Chrome on Windows 7, but this problem is also apparent when seen on Firefox. I just noticed this today and was wondering if it is a glitch that can be fixed. Thanks! —WP:PENGUIN · [ TALK ] 17:57, 5 October 2011 (UTC)

Already reported above. –Drilnoth (T/C) 18:04, 5 October 2011 (UTC)

General notice

  Resolved
 – already done

Given the scale of issues introduced by this new version of software, should we consider placing a general message suggesting readers head here to add issues they're experiencing? The Rambling Man (talk) 18:29, 5 October 2011 (UTC)

That sounds like a really good idea. --Kumioko (talk) 18:34, 5 October 2011 (UTC)
Already at the top of the watchlist. ---— Gadget850 (Ed) talk 18:51, 5 October 2011 (UTC)
Not sure if it's possible but probably worth re-asserting the message every so often because I usually ignore these things, so I guess many other readers do too. Since we have so many issues, it'd be worth re-iterating the fact that the site's broken but won't be broken for long... hopefully... The Rambling Man (talk) 19:08, 5 October 2011 (UTC)

table sort headers

  Resolved
 – fixed

As the whole header cell is now a button to sort the table, it has killed any header text that is linked. Links still show but don't direct. For example, try clicking on the US$ link on the tables in List of countries by GDP (nominal) per capita. With some header text entirely linked, and necessarily so in many cases, it would be good to be able to override this new feature.- J.Logan`t: 19:01, 5 October 2011 (UTC)

Please add to #The new Sorting breaks all tables arround Wikipedia which include two headers. ---— Gadget850 (Ed) talk 19:04, 5 October 2011 (UTC)
This is fixed now. --Catrope (talk) 19:39, 5 October 2011 (UTC)

Rollback WM

  Resolved
 – It's going far smoother than the 1.17 upgrade earlier this year, so we don't expect to roll back.

Just a question really, given the sudden swath of issues described in many sections above, is there any viability in rolling back to the previous version of Wikimedia software until such a time the major glitches are fixed? The Rambling Man (talk) 20:30, 5 October 2011 (UTC)

We do MW upgrades 1 to 3 times a year typically. They always have some teething problems that last a few days. However, as long as most of the content is usable for most readers, it is unlikely that a blanket rollback would happen. Developers are of course working to isolate the broken bits and either fix or revert the individual issues. Dragons flight (talk) 20:49, 5 October 2011 (UTC)
Cool, I'm certainly not trying to be difficult, it's just that I've never known Wikipedia to be so broken. Is there a timetable to correct the above issues, existing bugs etc? Sorry if I'm asking stupid questions but I am a mere user of Wikipedia and want to understand when it'll be as good as it was a few days ago. The Rambling Man (talk) 20:51, 5 October 2011 (UTC)
Based on past experience, most reader visible bugs should be cleaned up within a day or two. Bugs for editors may take a few more days if they are in Mediawiki's core, and up to a few weeks if they are in third party scripts (e.g. Huggle, Twinkle, etc.) since those depend on third-party developers that often don't have the time / resources of Wikimedia staff developers. A small fraction of the bugs may linger for a long time if they are either hard to diagnose (e.g. intermittent failures, specific browser versions) or if they interact with other software in a way that makes them both hard to fix and impossible to rollback without breaking other things. It is hard to put a time table on any specific bug, but most things should be back to normal within several days. Dragons flight (talk) 21:59, 5 October 2011 (UTC)
Ok, that's appreciated. I know how it goes, I'm in the mainstream industry, but where I am, we do tend to do a lot more beta testing with real end-users. Perhaps I've missed other WM updates where similar has occurred, but with a website that attracts the traffic we do, a lot of these bugs should have been ironed out before the update. But then all project managers would say that, wouldn't that? The Rambling Man (talk) 22:03, 5 October 2011 (UTC)

Just out of interest, can anyone link me to the discussions about issues when we moved to WM 1.17, just as a comparison, to see the number and nature of issues? Cheers. The Rambling Man (talk) 22:13, 5 October 2011 (UTC)

Wikipedia:Village pump (technical)/Archive 85#MediaWiki 1.17 release is tomorrow (hopefully) would be a good starting point. Edokter (talk) — 22:22, 5 October 2011 (UTC)

logs

  Resolved
 – invalid
Extended content

I swear whenever a page gets moved, the log gets noted in the page that gets moved as well as the redirect. Looking at the history of Soda and candy eruption, there is no log of moves unless you look at the reirects only. Have I missed something? See here and here. Simply south...... creating lakes for 5 years 20:57, 5 October 2011 (UTC)

The move is noted in the edit summary on both pages, but there's only one log entry at the original destination. That was the case before too. Reach Out to the Truth 03:08, 6 October 2011 (UTC)

Signpost

With all the other issues happening, I realize this is minor. But has anyone else noticed that EdwardsBot didn't deliver the Signpost this week? Maile66 (talk) 23:51, 5 October 2011 (UTC)

Yeah, I noticed that as well -- seems like all the bots are put on hold ATM, pending the complete roll out of 1.18. Sp33dyphil "Ad astra" 02:07, 6 October 2011 (UTC)
  Resolved
EdwardsBot has delivered the Signpost.

Who moved the feedback section?

Reported: … at top of page in colognblue skin☠MarkAHershberger☣ (talk) 05:11, 6 October 2011 (UTC)

Extended content

I use the Cologne Blue skin for WP. Yesterday, I think it was, the useless RATE THIS PAGE gibberish started appearing at the TOP of the page instead of the bottom, in all its real estate hogging glory. Where was the consensus for this?!? Or how does this bug get undone??? Carrite (talk) 04:24, 6 October 2011 (UTC)

Amazon Elastic Compute Cloud cache of WP

In this arstechnica piece on the Kindle Fire they mention an X-ray read-ahead cache of related Wikipedia content. Does this mean that we should expect Amazon Elastic Compute Cloud to be caching WP from now on? If so, does it open up any new performance possibilities? Just wondering... LeadSongDog come howl! 18:41, 29 September 2011 (UTC)

Doesn't seem to affect us since the article suggest they only pre-cache things for the kindle. Wikipedia is already cached quite heavily from Wikimedia servers. Bawolff (talk) 01:59, 7 October 2011 (UTC)

I think the secure login link on the login page needs to be changed: it gives https://secure.wikimedia.org/wikipedia/en/wiki/Special:UserLogin.

However, when switching between wikipedias (interwiki), it goes to a simple https://xx.wikipedia.org --- and therefore logs me out. Choyoołʼįįhí:Seb az86556 > haneʼ 04:10, 1 October 2011 (UTC)

The cross wiki problem is a known issue. I have changed the Login page to point to the new urls. There is one snag, I don't think there is a replacement atm for the "{{#ifeq: {{SERVERNAME}} | secure.wikimedia.org" trick that we used to change the content between secure and insecure servers. If anyone has an idea for a replacement, please do share. —TheDJ (talkcontribs) 11:11, 1 October 2011 (UTC)
That line is added dynamically in MediaWiki:Loginend when a user is not on secure.wikimedia.org. Unfortunately, there is no magic word to query the used protocol (there is $wgProto, but no {{SERVERPROTO}}), meaning that check will have to be removed. Edokter (talk) — 11:15, 1 October 2011 (UTC)
bugzilla:31293 proposes to simply split up the user message for the Special:UserLogin page from the server side. That would fix this specific problem. —TheDJ (talkcontribs) 22:20, 1 October 2011 (UTC)
Great. I hope this will be done for all wikipedias. Choyoołʼįįhí:Seb az86556 > haneʼ 02:35, 2 October 2011 (UTC)

my secure login is no longer available via the icon, if anyone understands this occurance then please explain Drift chambers (talk) 09:59, 6 October 2011 (UTC)

Looks like we've been upgraded

MW 1.18 is apparently here (Special:Version). There are a few teething issues (sidebar not collapsible, no edit toolbar... looks like ResourceLoader is dead). — This, that, and the other (talk) 00:31, 5 October 2011 (UTC)

Scripts are back. The Twinkle drop-down menu is now broken, so that will need to be fixed... aside from that, all seems to be OK. — This, that, and the other (talk) 00:34, 5 October 2011 (UTC)
Closing AfDs, the text in the "closing" and "relist" boxes is now like this and straining the eyes, is that a temporary thing I hope? - The Bushranger One ping only 00:37, 5 October 2011 (UTC)
Not only is the Twinkle menu visually broken, but the scripts themselves are still having issues. I just now got a "User talk page modification: Unknown error received from API while saving page" error. Regards, Orange Suede Sofa (talk) 00:38, 5 October 2011 (UTC)
And Twinkle is occasionally 'hanging' while closing AfDs. Sigh! - The Bushranger One ping only 00:52, 5 October 2011 (UTC)
I notice all the tabs and sidebar links are underlined now. It looks pretty ugly (but I still much prefer the underlines in the body which is why I have that setting.). Also, the 'm' and 'b' notes on the watchlist aren't bold any more. ♫ Melodia Chaconne ♫ (talk) 01:55, 5 October 2011 (UTC)
My watchlist isn't working correctly. Multiple changes to the same page are supposed to be collapsed together, but it's not working... Also, there's an odd glitch with the down arrow next to Twinkle, the page menu gadget, etc. It's showing 5 times in a row, the same distance apart, and the text goes in front of it. Might either of these things have to do with the new version? →Dynamic|cimanyD← (contact me) 02:47, 5 October 2011 (UTC)

In the "user contributions" list the ability to filter by namespace is gone.[16]   Will Beback  talk  01:55, 5 October 2011 (UTC)

This was reported on bugzilla:31197. Helder 02:05, 5 October 2011 (UTC)
Citations in {{reflist|group=upper-alpha}} are broken, see South American dreadnought race#Footnotes. There's also an extra drop-down arrow over "TW" in the top right corner (presumably only for users of Twinkle). Ed [talk] [majestic titan] 02:46, 5 October 2011 (UTC)
See #bug in #tag:ref parser function. I did fix one use that had quotes around "upper-alpha"— see Help:Cite link labels. ---— Gadget850 (Ed) talk 09:51, 5 October 2011 (UTC)
Who ever fixed the ref-note glitch -- Thank you!!! -- Gwillhickers (talk) 07:06, 6 October 2011 (UTC)
I have the same arrow problem, except since I use this I have several extra arrows. →Dynamic|cimanyD← (contact me) 02:50, 5 October 2011 (UTC)
User talk:72.27.85.119 is also borked in a really weird way... Hersfold (t/a/c) 03:05, 5 October 2011 (UTC)
Yet a permanent link to the current revision of the page works just fine. →Dynamic|cimanyD← (contact me) 03:08, 5 October 2011 (UTC)

One good thing I noticed is the message you see when you watch/unwatch a page now slides down smoothly, instead of just appearing. →Dynamic|cimanyD← (contact me) 03:10, 5 October 2011 (UTC)

Naturally, of course, no one has noticed or cared about this yet since it's an improvement, not a problem. →Dynamic|cimanyD← (contact me) 20:06, 6 October 2011 (UTC)

The letter tags for minor edits ("m") and bot edits ("b") in page histories and my watchlist are no longer showing in bold type for me. —{|Retro00064|☎talk|✍contribs|} 03:14, 5 October 2011 (UTC)

Wasn't it possible before to click on links in diffs? That no longer works. Orange Suede Sofa (talk) 03:18, 5 October 2011 (UTC)

What do you mean? —{|Retro00064|☎talk|✍contribs|} 03:28, 5 October 2011 (UTC)
Wikilinks in the body of diffs used to be displayed with markup (e.g. [[foo]] instead of foo), but were still clickable. Now they are still displayed with markup but are not clickable. Or at least I think it used to work that way. I remember it being very useful when verifying sources directly from a diff. Orange Suede Sofa (talk) 03:46, 5 October 2011 (UTC)
Unless there's a local site hack or gadget that I'm overlooking, I don't think that was ever the case. You can check a diff on dewiki (which is still running 1.17) for comparison.--Eloquence* 04:08, 5 October 2011 (UTC)
Oh, you're right. I have "Improved diff view" checked in my gadget preferences. I'm assuming that is it. Ok, then... that's broken. Regards, Orange Suede Sofa (talk) 04:13, 5 October 2011 (UTC)
solved☠MarkAHershberger☣ (talk) 04:26, 6 October 2011 (UTC)
I came over here hoping to find out why, since approximately 00:00 UTC today, any page looks as if I hadn't logged in or my monobook skin disappeared. Tabs which were at top of page (history, edit, talk, etc.) are gone, though a search down the page found the functions themselves. The entire left margin (with the Wikipedia globe and links to the Main Page and all sorts of other functions) is gone. It was OK on Wikimedia Commons for about an hour then, but the same thing had gone whacko there when I returned after a few hours offline. If an upgrade caused this, I hope there's a downgrade very soon... – Athaenara 07:49, 5 October 2011 (UTC)

Is this also the reason why the website keeps crashing? I've just wasted an hour trying to edit one page. DrKiernan (talk) 08:32, 5 October 2011 (UTC)

Yes, seems so. It works in Mozilla but not Explorer. DrKiernan (talk) 09:06, 5 October 2011 (UTC)

Invisible block: should blocked user be told automatically how to get third party review?

I was told on my talk page that my bot account had been blocked. However, I can't see any indication of the block. I don't want to try much testing in case it gets this account blocked too.

I'm not seeking opinions on the merits of the block. I'm sure that will all be sorted out. I'm merely seeking technical help on the following:

  • why is the block invisible to me at the blocked account (Lightbot) even if I log in as that account?
  • I understand blocked users are told how to get a third party review. Why is that also not visible to me?
  • what is the method of getting third party review?

Thanks. Lightmouse (talk) 01:07, 5 October 2011 (UTC)

http://en.wikipedia.org/w/index.php?title=Special:Log/block&page=User%3ALightbot shows the block fairly clearly. ΔT The only constant 01:10, 5 October 2011 (UTC)

I didn't know of that page. Is there a particular reason why blocks don't show at user pages? Lightmouse (talk) 01:45, 5 October 2011 (UTC)

it only shows in the block log and on the contribs page. ΔT The only constant 01:46, 5 October 2011 (UTC)
The log exists for other users to see the block. It is prepended to the user's contributions page, and to their userpage if it is empty. The user who is blocked will receive this message when they try to edit, explaining who blocked them, when, for how long, why, as well as giving four ways to appeal. Prodego talk 01:49, 5 October 2011 (UTC)

Interesting. Why don't we tell users before they try to edit? Lightmouse (talk) 01:53, 5 October 2011 (UTC)

Tell them what before they try to edit, exactly? That they're blocked? KillerChihuahua?!?Advice 03:11, 5 October 2011 (UTC)

Yes. But more importantly, how to get unblocked. The same information before the attempted edit as after. 09:42, 6 October 2011 (UTC)

Preventing an account from becoming autoconfirmed

Is there any way to prevent one's account from ever becoming autoconfirmed? I'd like to do that with my alternative account. A. di M.plédréachtaí 13:30, 5 October 2011 (UTC)

By not exceeding 9 edits. But to what end? –xenotalk 13:35, 5 October 2011 (UTC)
I actually have the same problem. I answer help desk questions and often log in to a non-confirmed alternative account to see what new users see. A solution to this particular problem would be to make two alternative accounts and make sure one of them never reaches 10 edits. If your concern is that somebody accesses your alternative account after use on a public computer and uses it to do inappropriate things only autoconfirmed users can do then I wouldn't worry about it. It's an unlikely scenario, it's limited how much damage they could do, and we already deal with scores of autoconfirmed bad apples. PrimeHunter (talk) 14:02, 5 October 2011 (UTC)
You can also set an edit filter to revoke autoconfirmed on the account, I believe. The edit filter could be one-use only. T. Canens (talk) 15:25, 5 October 2011 (UTC)
Alternatively, if you don't intend to ever edit from your alternate account but only view, but want to be sure you don't accidentally slip up and edit using it out of habit, you could ask the admins to simply indef-block the alternate account. - The Bushranger One ping only 19:36, 5 October 2011 (UTC)
What will result in wp:autoblock Bulwersator (talk) 13:25, 6 October 2011 (UTC)
Autoblock can be disabled, but if the purpose of having a non-autoconfirmed account is to "experience what new users experience", being blocked probably wont help - unless you want to experience what it is like for a new users to be blocked. –xenotalk 13:30, 6 October 2011 (UTC)

Strange changes

  Resolved
 – duplicate of #"m" and "b" edit indicators aren't bold
  Resolved
 – duplicate of #Namespace option on contribution history
Extended content

On user contributions pages, I noticed today that the m for minor was no longer bold and has an ellipsis underneath, and also I can nolonger select a namespace to view the contributions in. Has anybody else noticed this, and if so, what is going on? Rcsprinter (rap) 15:40, 5 October 2011 (UTC)

Workaround is to add the following to your css declarations:
/* Remove the distracting underlines from N, m, b in watchlist, make 'em bold */
 
abbr.newpage {border:0; font-weight: bold;}
abbr.botedit {border:0; font-weight: bold;}
abbr.minoredit {border:0; font-weight: bold;}
With credit to Redvers and SoWhy. –xenotalk 15:44, 5 October 2011 (UTC)
(edit conflict) Both topics are raised above. - David Biddulph (talk) 15:45, 5 October 2011 (UTC)
I am not seeing where. I hope that namespace search is restored quickly, I don't particularly care to learn css right now. --Piotr Konieczny aka Prokonsul Piotrus| talk to me 19:33, 5 October 2011 (UTC)
No need to learn - just cuta nd past the above into the css page. Thanks for that, that works even better than the bolding script above. I can see clearly now, the dots are gone! - The Bushranger One ping only 19:38, 5 October 2011 (UTC)
The "ellipsis" underneath the current unmodified display is your browser's default decoration for text associated with an abbreviation note (hence the "abbr" in the above css code), which looks a bit strange for only one character. On Firefox, for example, you get a question mark cursor if you hover over one of those letters and after a short delay an explanation pops up. --Mirokado (talk) 13:02, 6 October 2011 (UTC)

The Wikipedia Loves Libraries banner

The Wikipedia Loves Libraries banner is superimposing over other things at the tops of pages, such as geographic co-ordinates and the link to Help:User contributions on User pages. I'm using Firefox 7. The Mark of the Beast (talk) 22:17, 5 October 2011 (UTC)

Where are you seeing this banner? Is it rotating along with the maintenance message (in which case it might just not be showing for me right now?) or is it only in particular places? --brion (talk) 01:31, 6 October 2011 (UTC)
I see the WLL banner, but I'm not seeing it over other things. Maybe you have a gadget or other custom JS interfering? — ☠MarkAHershberger☣ (talk) 05:19, 6 October 2011 (UTC)
I am not using any gadgets, and I have not customized anything. The Mark of the Beast (talk) 05:41, 6 October 2011 (UTC)
What pages are you seeing the problem on? — ☠MarkAHershberger☣ (talk) 05:46, 6 October 2011 (UTC)
As I said above, on pages with geographic coordinates, and User contribution pages. Plus I keep hiding the damn banner and it keeps coming back anyway. The Mark of the Beast (talk) 06:45, 6 October 2011 (UTC)
Would you mind linking to a couple of those specific pages that you actually see a problem on? Also, what browser & browser version are you running? --brion (talk) 23:39, 6 October 2011 (UTC)
I would, but I've hidden the banner and now it seems to be actually hidden. If it shows up again, I will do that. I'm using Firefox 7. The Mark of the Beast (talk) 05:46, 7 October 2011 (UTC)
Funny, now that I've said that, it's back. See [17] and [18]. The Mark of the Beast (talk) 05:47, 7 October 2011 (UTC)

Namespace in Contributions

  Resolved
 – duplicate of #Namespace option on contribution history
Extended content

Selection by namespace has disappeared from Special:Contributions. Please re-instate instantly. There is a tick box labelled "Deleted only" - what is it supposed to do? — RHaworth (talk · contribs) 00:28, 6 October 2011 (UTC)

Seconded - what happened? This is somewhat essential for admins! ThanksSkier Dude (talk) 00:35, 6 October 2011 (UTC)

See #Namespace option on contribution history above. --brion (talk) 01:32, 6 October 2011 (UTC)

Someone please add a notice on top of Help:User contributions: there is a slight chance some people would look there before posting here. — AlexSm 01:57, 6 October 2011 (UTC)
A short notice in MediaWiki:Sp-contributions-explain would also help to avoid confusion for a lot of people. — AlexSm 02:17, 6 October 2011 (UTC)
Good idea.   Done Anomie 03:36, 6 October 2011 (UTC)
The box for deleted-only contributions was already there, but I've never figured out how it's helpful. Nyttend (talk) 04:09, 6 October 2011 (UTC)
The "deleted-only contributions" is for edits where RevDel was used. עוד מישהו Od Mishehu 08:59, 6 October 2011 (UTC)


Is using an ombox really necessary? That just makes it very distracting. I suggest changing it to something less attention-grabbing like "Note: The option to filter the contributions list by namespace has been removed in the upgrade to MediaWiki 1.18. Discussion is ongoing at Wikipedia:Village pump (technical)#Namespace option on contribution history and T33197." →Στc. 04:26, 6 October 2011 (UTC)

No such special page

Reported: no such special page -- MarkAHershberger 15:13, 6 October 2011 (UTC)

Extended content

Clicking on a Recent changes entry with the URL http://en.wikipedia.org/w/index.php?title=Sameera_Mallawarachchi&curid=33321830&action=history caused an unexpected response. The link is to an article I had just deleted. Strangely, the link shows a Wikipedia page with a title of Error, an article-like title of No such special page and content of You have requested a special page that is not recognized by Wikipedia. A list of all recognized special pages may be found at Special:Specialpages. If the &curid field is removed, it responds more sensibly showing deletion information.

I don't know if this is new behavior since the Wikimedia upgrade, but it seems a bit ungraceful. Either it should ignore the &curid= and report the page deletion, or say that the specific history items is not available. Naturally it should not mention anything about a Special: namespace operation. —EncMstr (talk) 09:22, 6 October 2011 (UTC)

How unusual; I've never seen such a thing before. Nyttend (talk) 10:50, 6 October 2011 (UTC)

subpage for new code deployment

It looks like there are going to be a lot of issues that need to be worked through. Should we create a separate subpage for reporting issues related to this deployment? John Vandenberg (chat) 09:37, 6 October 2011 (UTC)

Tab bars and apostrophes

Can someone maybe help out with a bug in a template? I've been rolling out {{start tab}} as a generic implementation for all these tab bars that are so popular with WikiProjects. See the one on Wikipedia:WikiProject Israel for an example: one central tab template which modifies its appearance automatically based on the page it's transcluded to. The problem is that it doesn't seem to work properly if the page title has an apostrophe in it: see Wikipedia:Romanian Wikipedians' notice board, where the current tab isn't styled differently. {{start tab}} draws its detection logic from {{tab}}, so that'd be the most likely place to look for a fix. Any thoughts? Chris Cunningham (user:thumperward) - talk 09:48, 6 October 2011 (UTC)

{{FULLPAGENAME}} outputs the apostrophe as "&#39;" (the HTML numeric entity for the apostrophe character). In the actual HTML output, Tidy seems to change this to the apostrophe character, but that hasn't happened by the time it gets to your {{#ifeq:}}. If you pass the page name in your {{start tab}} invocation with apostrophes replaced by "&#39;", it should work. BTW, the same happens for double-quotes (by "&#34;") and ampersands (by "&#38;"). Anomie 11:07, 6 October 2011 (UTC)
Awesome: works perfectly. Meh, entities. Chris Cunningham (user:thumperward) - talk 13:19, 6 October 2011 (UTC)

Problem with Internet Explorer

Reported: 1.18 Causes IE8 to crashMarkAHershberger 14:47, 6 October 2011 (UTC)

Extended content

When opening a page on the English wikipedia with Internet Explorer, an error message appears and the page has to be re-opened. When making changes and visualising them, the changes are lost, so that I had to make the changes again and to save them without visualising them.

Please solve this issue. --Réginald alias Meneerke bloem (To reply) 12:24, 6 October 2011 (UTC)

[[File:foo.ogg|noicon]]: noicon ignored

Reported: Noicon parameter for audio files is brokenMarkAHershberger 14:47, 6 October 2011 (UTC)

Extended content
It seems the parameter |noicon in the code [[File:Accordion_chords-01.ogg|noicon|right]] is no longer observed; vide:
noicon

The template {{Listen/core}} issues this parameter and ignoring it changes the appearance of {{Listen}}. What caused this change and how can it be reverted? -- Michael Bednarek (talk) 13:27, 6 October 2011 (UTC)

Fix is queued up in trunk: mw:Special:Code/MediaWiki/99167. --brion (talk) 23:35, 6 October 2011 (UTC)

Yet another problem caused by the archiving — something wrong in the software about section editing

Hello,

Following of this discussion.

The discussions get stupidly archived by date, even when they are still very recent and very relevant. And so the links to the discussions get broken. And the archived page tells not to edit the page, but to start a new discussion if one wants to say something.

On the page Wikipedia talk:Manual of Style, there was the section Making MOS:ENDASH happen. I clicked its edit link. I landed into the edit form of the section Specific–vague axis in choosing article titles, a completely different section !

What happened ? The section I intended to edit was still on the main page, but someone had archived some other section(s), and so my intended section's order number had changed. And the software, not very clever, uses only the section's order number to identify a section, and believes this is sufficient.

There is something wrong in the software about section editing. The software has to handle sections better than that !

--Nnemo (talk) 14:31, 6 October 2011 (UTC)

liquidthreads problem at en.wiktionary

Reported: Bugzilla:31251MarkAHershberger 14:41, 6 October 2011 (UTC)

Extended content

Hi, on en.wiktionary, I see "my new messages (1)" in bold, but when I click on it, it says "There are no new messages for you.".

Is there a way to unbold the link, since there aren't actually any messages there? -- Liliana-60 (talk) 14:34, 6 October 2011 (UTC)

Tables now going "off screen" (and no scroll bar in Internet Explorer)

Reported: wider table columns causing problems in IEMarkAHershberger 17:31, 6 October 2011 (UTC)

Extended content

Something has happened to the sortable table columns. They have become much wider, leading the table at Snooker world ranking points 2011/2012 to go off the side of the page. In Firefox this isn't so bad because you can at least scroll along. However, in Internet Explorer 8 it is impossible to even scroll along meaning we lose half the table. I can see the whole chart in IE8 if I set my scale to 75% (on a 1024x768 screen), but at 100% some of the columns off the screen and it doesn't have a scroll bar at the bottom to move along like in Firefox. I'm sure our table isn't the only one affected on Wikipedia if the columns have been widened (which I didn't think was necessary to be honest because the chart looked better when it was fully on the page), but by not being able to scroll along in IE it has become virtually useless. It would be great if someone could address the IE issue. Betty Logan (talk) 17:14, 6 October 2011 (UTC)

You could wrap it in an overflow div, or remove the nine empty columns. — Bility (talk) 17:16, 6 October 2011 (UTC)

Diff Categorizer Gadget

We are looking for some technical feedback on a gadget for letting users categorize edits. Please have a look at the gadget proposals page for details. We will be grateful for any comments. Also look at the user instructions. Fikonträd (talk) 17:35, 6 October 2011 (UTC)

Block log entry behaving erratically

  Resolved
 – Fixed by bugzilla:31352.
Extended content

This entry of an 2008 block (presumably for notation purposes) shows an expiry time in 2011. Moreover, the expiry time shown seems to be actually the current time - it actually changes over time. T. Canens (talk) 17:41, 6 October 2011 (UTC)

Parentheses pipe shortcut doesn't work in <ref>

Wikipedia allows the shortcut of simplifying the piping of wikilinks containing parentheses, for example, ''[[Nature (journal)|Nature]]'' (Nature) can be simplified to ''[[Nature (journal)|]]'' (Nature). This works inside templates, including most importantly the various {{cite}} templates.

But it doesn't work when nested in a <ref>. For example, the reference

{{cite journal |author=Joe Shmoe |year=2011 |title=Piped Wikilinks with parentheses |journal=[[Nature (journal)|]] |volume=333 |issue=1 |pages=551–552}}

displays correctly in running text as

Joe Shmoe (2011). "Piped Wikilinks with parentheses". Nature. 333 (1): 551–552.

and incorrectly as a footnote.[1]

Bare ''[[Nature (journal)|Nature]]'' displays OK in a <ref>[2] — but ''[[Nature (journal)|]]'' displays in a <ref> as if everything but the apostrophes were nested in an <nowiki></nowiki>.[3]

Is this a known issue and are there any plans to fix it? —Anomalocaris (talk) 18:43, 6 October 2011 (UTC)

  1. ^ Joe Shmoe (2011). "Piped Wikilinks with parentheses". [[Nature (journal)|]]. 333 (1): 551–552.
  2. ^ Nature
  3. ^ [[Nature (journal)|]]


This is a very old known bug: bugzilla:2700. Not very high priority, apparently. :) --brion (talk) 18:51, 6 October 2011 (UTC)

Section erroneously hidden

The section #Improved diff view not working has been hidden in side the extended content of the previous section, #Wikilinks in diffs no longer clickable, which is listed as fixed. However, they are different (no pun intended) problems, and improved diff view is still not working. --Stfg (talk) 18:58, 6 October 2011 (UTC)

Now dealt with. Thanks, Optimist on the run. --Stfg (talk) 21:09, 6 October 2011 (UTC)

autoblocking weirdness

See User talk:Esoglou#Unblock-auto. User was somehow autoblocked for the mere act of logging into an old sock account to check its watchlist. Hasn't edited with it since late 2009. Beeblebrox (talk) 23:22, 6 October 2011 (UTC)

Different formatting of categories?

It looks like the categories box at the bottom of most pages has been reformatted to include more space between the category names. Where was this discussed or where would it be discussed? --Metropolitan90 (talk) 05:20, 5 October 2011 (UTC)

Yes, I believe it was a result of the MW 1.18 rollout (see above). Not sure where you would go to discuss it, someone else might, though. Jenks24 (talk) 05:23, 5 October 2011 (UTC)
...I just noticed that now that you mentioned it. It looks horrible. So does text in the close/relist AfD boxes being this size. - The Bushranger One ping only 05:44, 5 October 2011 (UTC)
Looks good - more modern and Vector-ish, in my opinion - unlike the old category-link style that was left over from Monobook. — This, that, and the other (talk) 05:52, 5 October 2011 (UTC)
That's all well and good - except I have (and prefer) Monobook, and would like to at least have the option to have the old style display. - The Bushranger One ping only 06:03, 5 October 2011 (UTC)
It does look terrible on Monobook. "More modern" does not necessarily mean "better". --jpgordon::==( o ) 06:13, 5 October 2011 (UTC)
Yes, I can imagine that it would jar with the small, sharp style of Monobook. Perhaps we can change it locally through MediaWiki:monobook.css? — This, that, and the other (talk) 06:36, 5 October 2011 (UTC)
I use the classic monobook as well, and yes, the categories look weird. The tiny inline icon that used to appear after an external link is now missing too... all I see there now, is a blank whitespace.  -- WikHead (talk) 07:45, 5 October 2011 (UTC)

Don't call "Monobook" classic, because "Classic" is a different skin... and the categories look horrible there too. DS (talk) 23:07, 5 October 2011 (UTC)

It looks terrible in Monobook, so I just now switched to Vector, and I think it looks just as bad in Vector. --NSH001 (talk) 09:09, 5 October 2011 (UTC)
It looks like it was done in rev:92054 (bugzilla:12261), due to the change in how they are rendered. Peachey88 (T · C) 08:02, 5 October 2011 (UTC)
Blech! I agree that the categories look quite terrible now. --Philosopher Let us reason together. 19:35, 5 October 2011 (UTC)


Anyone who doesn't like the new styles should be able to get the old ones back pretty easily. Let me know if you need help setting up your CSS for this. — ☠MarkAHershberger☣ (talk) 04:50, 6 October 2011 (UTC)
This category issue is clearly associated with the move to Mediawiki 1.18. It is almost certainly a BUG, probably technical but at the least aesthetic. Even if 1.18 somehow fixed a CSS rendering issue and is now rendering the categories "correctly", the skins were designed with the old rendering in mind. Regardless of the technical question of the correctness of the new rendering, the new look for the categories is terrible. There's simply way too much space being used for the category separators. (I am using Vector and haven't checked Monobook.) This should be investigated and the look should revert to the previous spacing. Jason Quinn (talk) 00:14, 7 October 2011 (UTC)
See Wikipedia:Village pump (technical)/Archive 94#Categories are surrounded by much more space, and they don't wrap from line to line for a chunk of CSS to restore something like the old behaviour. --Redrose64 (talk) 11:35, 7 October 2011 (UTC)

Is there any way to have a button, tab or a link that I could customize to add a certain text? I'd like to be able to "one click" my default welcome templates, and messages like this. Any suggestions? --Piotr Konieczny aka Prokonsul Piotrus| talk to me 17:51, 5 October 2011 (UTC)

You can do this with Twinkle, but you probably want to wait until issues with the update shake out. ---— Gadget850 (Ed) talk 18:39, 5 October 2011 (UTC)
Take a look at meta:User:Krinkle/Scripts/InsertWikiEditorButton. Helder 15:01, 7 October 2011 (UTC)

We need to do a better job of this

I realize that developing new changes to a widely used application like this are not at all easy and I don't mean to be overly critical but after reviewing the growing list of problems with the new release and the fact that most of us didn't even know it was coming I think we need to do a better job.

Several of the problems identified would have been obvious in some basic comparison testing and many of the common functions would have also been obvious if done (like all the problems with Twinkle). I realize that its impossible to compensate for every problem and WP has limited support of add ons like twinkle but some of these are really really obvious.

I recommend when we do updates in the future we do a couple things differently:

  1. Put a notice on this village pump of the upcoming update
  2. Establish a schedule of the updates in a public place like this page
  3. Do some quality testing with some of the widely used tools like Twinkle and let the developers of those apps know that there are problems they need to fix.

--Kumioko (talk) 18:31, 5 October 2011 (UTC)

I know it's a minor part of Wikipedia but yes, some of our featured material (lists) have now been broken with the update, these feature once a week on the main page. This sort of thing should have been tested/advertised before being wholesale rolled out. RIght now we don't know whether to stick (not fix) or twist (try to fix hundreds of lists)... advice please. The Rambling Man (talk) 18:35, 5 October 2011 (UTC)
Posting a schedule in a public place was done weeks ago. --Catrope (talk) 19:22, 5 October 2011 (UTC)
Notices on the Village Pump / Template:CENT would probably have a lot more outreach than the WMF blog. Headbomb {talk / contribs / physics / books} 19:37, 5 October 2011 (UTC)
You mean like announcing it on VPT three weeks ago, Wikipedia:Village_pump_(technical)/Archive_93#MediaWiki_1.18_deployment_schedule? Dragons flight (talk) 19:42, 5 October 2011 (UTC)
Yep, missed that, not a watcher of the Village Pump. Didn't expect Wikipedia to be wrecked by a software "upgrade". Now I know what it was like when people moved to Windows Vista. Ouch. The Rambling Man (talk) 19:47, 5 October 2011 (UTC)
Or also, Wikipedia:Wikipedia_Signpost/2011-09-19/Technology_report, from just over two weeks before deployment. Dragons flight (talk) 19:58, 5 October 2011 (UTC)
So if you didn't watch the Village Pump or subscribe to Signpost, then you have no idea why Wikipedia is broken. Don't forget that this website is visited by users who don't even know that Village Pump or Signpost exist. We're just telling those who rolled out the bug-ridden version of Wikimedia software that those who are familiar with the site had no real clue it was happening and that it's broken a lot of stuff. Don't take it personally, just tell us when it'll all be fixed. The Rambling Man (talk) 20:09, 5 October 2011 (UTC)
(e/c) That is interesting, but how many regular Wikipedians read that? Perhaps a general notice for all readers, including anon IPs would be useful (if it doesn't already exist) to say that Wikipedia is currently broken and give normal users an idea when it will return to a useful state... The Rambling Man (talk) 19:38, 5 October 2011 (UTC)
(e/c)I was about to say the same thing. A general notice a week before linking to that blog and a page for problems to be listed would have been very useful I would think. --Kumioko (talk) 19:42, 5 October 2011 (UTC)
Were the developers/maintainers of tools like Twinkle and CSDHelper warned of the change and given access to an advance version to test with? JohnCD (talk) 22:11, 5 October 2011 (UTC)
Probably not directly, but it can be presumed that they have VPT on their watchlist. Remember that en.wiki is probably one of the most complex and intricate projects, and some bugs just need to be found in the wild. I appreciate all the hard work that went into this release and accept the growing pains as necessary... –xenotalk 22:36, 5 October 2011 (UTC)
I guess this is the short term message. As Xeno says, hats off to the people that do all this work, both before and in particular after each version's rollout. Just by looking at the bugzilla reports of things I care about, it's clear to see that people are working flat out to fix the issues. I for one am very grateful. —WFC00:21, 6 October 2011 (UTC)
For what are you grateful? That the developers work hard to eliminate problems they themselves introduced? If so you're very easily pleased. Malleus Fatuorum 01:56, 6 October 2011 (UTC)
Were there any upsides to this "upgrade"? All I can see at the moment is the things that don't work. Is anything actually better as a result? I might feel a little better about this if there was. --John (talk) 06:10, 6 October 2011 (UTC)
See here and the full release notes. MER-C 11:38, 6 October 2011 (UTC)
An impressive list indeed. Perhaps the best course of action is to join the Italians for a few weeks, hoping that the devs will somehow switch back to 1.17. NVO (talk) 11:56, 6 October 2011 (UTC)

() Lines 8-11 of those relase notes (dated Oct 3):
8 THIS IS NOT A RELEASE YET
9
10 MediaWiki 1.18 is an alpha-quality branch and is not recommended for use in
11 production.
--Stfg (talk) 13:35, 6 October 2011 (UTC)

MediaWiki has always been deployed to Wikipedia before the software is released as a tarball (what the release notes cover). The blog post covering 1.18 deployment points to this pageMarkAHershberger 13:46, 6 October 2011 (UTC)
Just to clarify I know that its difficult to do a release like this and if it were just some problems with addons like twinkle that arent a part of the software or a few little bugs it would be fine. But when there are problems like breaking all the featured tables, screwing up the diff rendering, completely removing the ability to see namespace specific changes in contributions and a variety of others it would have been immediately obvious to any tester that these were problems if they would have done some simple comparisons. I also agree with the comments above that leaving the notice on the blog and Tecnhical village pump are noble but less than a 100 of our millions of users are likely to see either of those. I also agree with the comments that the developers are working hard to fix them but may of these should have never been put in production to begin with so now we have to live with the fallout. Personally I have a huge amount of edits and things I am trying to do in WP that now have to wait because it takes the page about 1-5 minutes to load a page every time I bring a new page up in edit mode, making it nearly useless and next to impossible to do a change. So whenever we get these problems fixed in the next couple weeks I will do the WikiProject United States Newsletter, continue working on project issues and contributing content. --Kumioko (talk) 14:57, 6 October 2011 (UTC)
Re “it would have been immediately obvious to any tester that these were problems if they would have done some simple comparisons”:
  • breaking all the featured tables
Agreed, but we do testing before deployment. How can we make sure all these template things are checked? (And, yes, I want to know how you think we could do it.)
  • screwing up the diff rendering
This was a gadget problem. If you want to make sure your gadgets are working, then you should help test.
  • completely removing the ability to see namespace specific changes in contributions
This was a policy decision to limit use of database resources. Obviously it should have had some more discussion with the community before hand, but it isn't a bug that we didn't know about.
MarkAHershberger 15:57, 6 October 2011 (UTC)
I've moved the discussion about how we could improve testing before deployment. — MarkAHershberger 19:10, 7 October 2011 (UTC)
There was a fair number of issues with the 1.17 deployment, including an initial failed deploy and, again, a number of issues with gadgets, scripts, and breakage of some features or some configurations. You can read all about it here and here if you'd like to experience some deja vu.
There are some factors that make deploys hard for us that are fairly unique to us (for example, we give users an unprecedented level of control over the site by enabling admins to change site scripts and CSS, and we have a very powerful but also fairly fragile framework for gadgets), there are some that aren't unique to us but that aren't easily resolvable either (we're running an unusually large and complex site without the commensurate level of staffing e.g. for QA automation and testing), and there are some that are very much resolvable and in our control. Let me focus on the last ones.
Firstly, our release cycle has slowed down to the point where we push out releases with a huge number of changes that have been sitting unused in our version control system for months. See the release notes for 1.18 and 1.17. We've about cut the time between releases in half from 1.16->1.17 to 1.17->1.18, which is good progress, but we really want to get to an even more frequent cadence to achieve continuous integration of smaller increments of features, so that issues can be detected at a more granular level and be resolved close to the time when they've been caused by a code change.
Second, our unit testing is slowly becoming useful. We've put quite a bit of effort into increased unit test coverage with PHPUnit and Qunit and a new continuous integration server that regularly runs those tests, and we're starting to be bump code immediately when it fails tests (which is standard practice, but not something we've actually been doing).
Third, for gadgets, we've been working on supporting centralized publication of gadgets through the ResourceLoader 2 framework. This will make it possible to tests gadgets more systematically, and we can more easily set up test wikis that subscribe to key, frequently used gadgets. We should also start pushing gadget developers to write unit tests, and integrate them with the CI process.
Fourth, we've implemented support for heterogeneous deployment through the new multiversion framework. This has allowed us to push 1.18 out much more gradually and systematically than 1.17, and to have multiple stable testing sites that run on the production infrastructure but run different versions of the code.
Fifth, we've been working on a dedicated OpenStack-based infrastructure for all our continuous integration needs, to replace the aging and insufficient prototype VMs. The new labs console is used to manage VM instances that will increasingly resemble the production environment. These VMs will be made available to all MediaWiki developers (staff and volunteers), enabling better testing as part of the development process. We should also be able to use them for automated integration testing at some future point.
Sixth, for cross-browser testing, we've deprecated our aging local VM infrastructure in the office, and began evaluating a number of commercial solutions, both for manual and automated testing. We'll likely use CrossBrowserTesting for manual testing, and some combination of a system like SauceLabs and our existing TestSwarm setup for automated integration testing.
Seventh, we don't have a QA department yet! As we're growing our engineering staff from almost nothing to a dedicated team of a couple of dozen people, this has been high on our priority list for a long time. We've been actively recruiting for a QA lead position, and we're also going to hire some contractors both for testing and community coordination. As you know, Mark Hershberger has been helping a lot already, and I think you can see the results of this with the 1.18 deployment (bugs are being triaged and tracked more effectively than 1.17). I also completely agree that we can do more effective community outreach.
In a nutshell, the combination of shorter release cycles and improved testing methodology should help us make releases incrementally less painful. The proof will be in the pudding, but I hope you can see that we're very actively improving our development and testing methodology all the time, and we always appreciate feedback on how to do better.--Eloquence* 19:33, 7 October 2011 (UTC)

{{REVISIONUSER}} not working all the sudden

Reported: Revision information not rendered in edit notices

See Template:Editnotices/Page/Pingan International Finance Center. It works when I preview the notice, but not when I try to edit the article. Checked another edit notice I made a while back and the same thing is happening there. Beeblebrox (talk) 20:53, 5 October 2011 (UTC)

If you mean it should the name of the user viewing it, then that has been fixed. It was always considered a bug. Let me dig up some details. ---— Gadget850 (Ed) talk 21:17, 5 October 2011 (UTC)
Well fixed, but not fixed. Yes, the prior behavior was considered a bug, but the current behavior for REVISIONUSER in an edit notice seems to be to display an empty string, e.g. top of this page. I would consider that to also be a bug. If I understand correctly, REVISIONUSER is supposed to display the name of the user who saved the topmost revision. Dragons flight (talk) 21:28, 5 October 2011 (UTC)
bugzilla:31398. Dragons flight (talk) 23:00, 5 October 2011 (UTC)
And T21006 started this. ---— Gadget850 (Ed) talk 23:53, 5 October 2011 (UTC)

Well I'm glad others have noticed this, because I spent an unfruitful hour last nigh trying to fix my talk page edit notice, thinking it was due to vandalism to my page. --Kudpung กุดผึ้ง (talk) 00:09, 6 October 2011 (UTC)

I added a notice to WP:Editnotice. ---— Gadget850 (Ed) talk 02:56, 6 October 2011 (UTC)
Just make sure it's not an editnotice.  Στc. 04:25, 6 October 2011 (UTC)
Some people were relying on the previous behaviour of {{REVISIONUSER}} - see, for example, MediaWiki talk:Histlegend#New tool for consideration. --Redrose64 (talk) 10:15, 6 October 2011 (UTC)
The same thing is happening to my editnotice for my userpage. HurricaneFan25 17:25, 6 October 2011 (UTC)

This has also broken all the policy and guideline editnotices, which use {{policy-guideline-editnotice}}, which uses {{REVISIONUSER}}. →Dynamic|cimanyD← (contact me) 20:03, 6 October 2011 (UTC)

Fixed by Σ. →Dynamic|cimanyD← (contact me) 12:11, 7 October 2011 (UTC)
  Resolved

{{resolved|invalid}}MarkAHershberger 17:31, 6 October 2011 (UTC)

See Jesse Jackson. There are no section edit links and I tried, but couldn't find a __NOEDITSECTION__ anywhere (I gave up after a while). Anyone else care to check? Thanks. howcheng {chat} 16:42, 6 October 2011 (UTC)

They are showing for me. –Drilnoth (T/C) 16:50, 6 October 2011 (UTC)
I see them too. Though one one occasion, I did see them disappear from another article as well, only to reappear after an edit. Edokter (talk) — 16:53, 6 October 2011 (UTC)
...and now they're back. I'll chalk it up to a fluke, then. howcheng {chat} 16:56, 6 October 2011 (UTC)
This happened to me yesterday as well, at WP:AN (Wikipedia talk:AN#Section edit buttons gone?). They re-appeared after I made another edit to the page. –xenotalk 17:01, 6 October 2011 (UTC)
This is happened again on 2PM page.--Lpmfx (talk) 19:11, 6 October 2011 (UTC)
This is not 'invalid' but I'm not sure how to reproduce. See below subsection. The edit links re-appear upon subsequent edits to the page. –xenotalk 19:20, 6 October 2011 (UTC)

We figured this one out at last. The secret to reproduce was something like this:

  • go to a printable page
  • if necessary, force an action=purge on the printable URL to make sure you've recached
  • go back to the non-printable view

Things like the printable view or viewing when you don't have edit permissions had forced the section edit links off, but because _most_ code paths no longer actually turn section edit links off it was colliding in the parser cache with the regular version.

The fix was to update things a little more so that when rendering for output that shouldn't have section edit links, we go ahead and prep them in the cached output but then don't include them at final output. This was already being used for, for instance, letting people with section edit links disabled not have to split the cache.

You might have to null-edit or action=purge affected pages, but they'll all clear themselves out over time. :) --brion (talk) 01:05, 8 October 2011 (UTC)


Can't edit sections of semiprotected articles

I now do not see an "edit" link next to the section titles of semiprotected articles, like Great Famine (Ireland), or Barack Obama, although I am logged in. If it matters, I am using the old-style Monobook view. Comet Tuttle (talk) 19:14, 6 October 2011 (UTC)

Combined with #No section edit links. –xenotalk 19:20, 6 October 2011 (UTC)

Category bloating

Hi. I noticed today that the categories have now got wider spacing between them. I just changed my skin to Cologne Blue and its buggered the alignment and looks ridiculous at the article now features further down with an ugly gap at the top. Especially for articles with a lot of categories and wiki links it looks horrendous (check out Austria in Cologne Blue for example).♦ Dr. Blofeld 21:02, 6 October 2011 (UTC)

I noticed that too, it seems like there's too much space between lines and between links, even on regular Vector skin. --Funandtrvl (talk) 21:14, 6 October 2011 (UTC)

Can we

  • a] Go back to the neat category alignment
  • b] Modify the Cologne blue skin to place categories and interwiki links at the foot of the page instead of at the top?

Dr. Blofeld 21:18, 6 October 2011 (UTC)

As I mentioned in the 1.18-bug section above, the new spacing - while the "no wrap" part is good - makes some articles look horrendous. - The Bushranger One ping only 22:22, 6 October 2011 (UTC)
See Wikipedia:Village pump (technical)/Archive 94#Categories are surrounded by much more space, and they don't wrap from line to line for a chunk of CSS to restore something like the old behaviour. --Redrose64 (talk) 11:38, 7 October 2011 (UTC)

Sorting update question

Under the old table sorting feature, the sorting image could be placed under the column heading with a line break (<br>). This was useful for keeping column width in large sortable tables as slim as possible. After the recent update the line break has no effect and the sorting image is now horizontally aligned to the right and vertically aligned to the center of the top row. Under the new update, is it possible to place the sorting image below the column heading? – Zntrip 22:24, 6 October 2011 (UTC)

Personally I prefer the fixed position of the user interface element actually. —TheDJ (talkcontribs) 12:34, 7 October 2011 (UTC)

But regardless, is there are way to change it? Like I said, it takes up a lot of space lengthwise that some tables can't spare; for such tables width usually isn't an issue, so there is a solution if the sorting image can be reoriented. Additionally, uniformity can still be maintained within the table. – Zntrip 20:31, 7 October 2011 (UTC)

Turning off fade-in/out effect

== watchlist javascript "upgrade" ==

When I expand/contract grouped watchlist edits now, I get a fancy fade-in/fade-out effect. This might look pretty, but it just seems to waste time. Is there a way I can shut this off? The un-pretty method was fine with me... --Fru1tbat (talk) 23:41, 6 October 2011 (UTC)

Try setting jQuery.fx.off in your common.js:
$.fx.off = true; //turn off all jQuery animations
AlexSm 01:25, 7 October 2011 (UTC)
Nifty. I was a bit annoyed by the watch/unwatch "slide" effect, this took care of that. Thanks! - The Bushranger One ping only 01:29, 7 October 2011 (UTC)
Sorry, my posts tend to come out a bit snarkier than I realize when I'm overworked. Thanks for the tip, though. Works like a charm! --Fru1tbat (talk) 01:58, 7 October 2011 (UTC)
Gadget time! Edokter (talk) — 11:52, 7 October 2011 (UTC)
Gadget added. —TheDJ (talkcontribs) 12:32, 7 October 2011 (UTC)
Also of interest: bugzilla:30401. Edokter (talk) — 15:26, 7 October 2011 (UTC)
Another option: mw:Snippets/Toggle animations. Helder 01:38, 8 October 2011 (UTC)

MathBot is broken

  Resolved

Among the multitude of borked bots on account of 1.18, User:MathBot is one I'd like to draw particular attention to - since it's important to the workings at AfD. The bot's owner doesn't appear to check in too often, is there anyone else who can fix at least the AfD-related parts of its operation? (It's creating the new AfD logs just fine, but claiming the old logs all have "0 open/0 closed/0 total" AfDs in them, which makes it troublesome to pinpoint AfDs that are "stale" and need closing/relisting in those logs). - The Bushranger One ping only 00:39, 7 October 2011 (UTC)

MathBot's AfD function is working now. Thanks to whoever fixed it. :) - The Bushranger One ping only 15:40, 7 October 2011 (UTC)

Banners

Just lately I'm being assaulted by banners, usually asking me to share my story for the 2011 fundraiser, but occasionally it's something about libraries or a survey about the mobile site, although these two are relatively infrequent. Clicking [X] helps for a moment, but the banner usually reappears a few pages later. Is there a way to block these? I didn't see anything under the gadgets menu in preferences. I may or may not have had something in my monobook.js that usually hid them, which may or may not have been disabled by the recent server upgrade. Thanks. --Bongwarrior (talk) 02:33, 7 October 2011 (UTC)

Its gotten so bad with all of the annoying global notices using CentralNotice that I added a rule about it to Adblock Plus. ΔT The only constant 02:38, 7 October 2011 (UTC)
Agree with Bongwarrior, but I don't think the upgrade caused this – I've never had any js to stop banners and it's never been this bad. The problem is I don't necessarily want to hide all notices, I just want them to stay away for more than a minute once I've clicked on the [X]. Jenks24 (talk) 02:44, 7 October 2011 (UTC)
I'd like to figure out a way to get rid of them too. I think, back when they had the Big Honking Banners with the pictures of people in them, there was a checkbox in Preferences that did it, but I can't find it now... - The Bushranger One ping only 02:43, 7 October 2011 (UTC)
There was. After the fundraiser ended, it was removed "until next year". Anomie 02:48, 7 October 2011 (UTC)
(EC x2) This is getting aggravating and i have come to the VPT to find if there is a solution. Before, when I closed it once, it would stay closed, but it seems to be ignoring that this time. hbdragon88 (talk) 02:44, 7 October 2011 (UTC)
Same problem for me. In fact a banner I've repeatedly clicked away popped up on this edit page. The behavior seems random and buggy. I've also noticed that when a banner you've clicked away reappears, if you navigate to another page without clicking it away again, chances are it won't be on the new page. They pop up randomly, it seems. Pfly (talk) 03:47, 7 October 2011 (UTC)
It looks like the problem is that they have the banner marked as "not a fundraising banner", but the 'hide' link is for hiding a fundraising banner. See here. It's probably intermittent because other banners in the rotation aren't broken. Anomie 04:01, 7 October 2011 (UTC)

I think I got it. Add div.siteNotice {display:none !important} under preferences > (your current) skin > Custom CSS. It appears to not work if you use HTTPS to browse the Wiki. hbdragon88 (talk) 04:12, 7 October 2011 (UTC)

It's not working for me. :( - The Bushranger One ping only 04:18, 7 October 2011 (UTC)
Hm. Removing the "!important" part of the code seems like it might have made it work. - The Bushranger One ping only 04:21, 7 October 2011 (UTC) ...and ironically, the post I made to say that, proved me wrong. Not working here. - The Bushranger One ping only 04:22, 7 October 2011 (UTC)
My apologies. Do this div.globalWrapper#div.column-content#div.siteNotice { display:none !important} that worked for me. hbdragon88 (talk) 05:22, 7 October 2011 (UTC)
It still popping up :(--♫Greatorangepumpkin♫Heyit's me 10:21, 7 October 2011 (UTC)
I personally use #centralNotice {display:none !important;}, and that seems to work nicely for me. Avicennasis @ 13:02, 9 Tishrei 5772 / 13:02, 7 October 2011 (UTC)

So I'm not the only one who can't get the banners to stay even if I click the [x], huh? Dear WMF, if you'd like to keep your volunteers happy, then MAKE SURE YOUR DISMISSABLE ADS CAN ACTUALLY BE DISMISSED. /ƒETCHCOMMS/ 04:40, 7 October 2011 (UTC)

I'm currently employing Δ's method of using Adblock Plus to squash them, and it seems to be working nicely. I didn't have any luck with Hbdragon88's solution, although I appreciate the suggestion (and it's eminently possible that I was doing something incorrectly when I tried it). Thanks. --Bongwarrior (talk) 07:24, 7 October 2011 (UTC)

It didn't work for me. Can you explain what you did?--♫Greatorangepumpkin♫Heyit's me 10:15, 7 October 2011 (UTC)
Click the Adblock Plus icon in the bottom corner of your browser -> click "Open blockable items". In the list that appears, look for any entries that mention banners. Right click on the entry and choose "Block this item", choose "Custom:url" and then "Add filter". I did this for two entries. --Bongwarrior (talk) 19:27, 7 October 2011 (UTC)

Shouldn't this be solved centrally, for all users? It is very annoying not being able to actually dismiss the banner. One should be able to remove such banners. Manxruler (talk) 09:47, 7 October 2011 (UTC)

There is also a problem that the "Get involved" button on the fundraiser banner appears in exactly the same place as a {{coord|display=title}}. Monobook, Firefox 3.6.23, Windows XP. --Redrose64 (talk) 10:27, 7 October 2011 (UTC)

Massive pain in the arse, to the point where I'm tempted to post a hail of troll-like abuse on whatever links appear that I'm asked to contribute toward. I'd be ecstatic if I never had to read another one of these stupid fucking notices again. Parrot of Doom 10:52, 7 October 2011 (UTC)

I'm with Parrot on this. What utter retard is responsible for this huge fuck-up? Lugnuts (talk) 12:12, 7 October 2011 (UTC)
The words, the words. Just because someone is not an editor right here, he is part of our community. Be more considerate with your words. —TheDJ (talkcontribs) 12:23, 7 October 2011 (UTC)
He? Well you've failed your equal opps too. 24hrs is more than enough time for this abortion to be fixed. Thanks. Lugnuts (talk) 12:24, 7 October 2011 (UTC)

When you have to keep click close on a banner like 40 times an hour this really is not good enough. Sort it out sombody please..♦ Dr. Blofeld 12:59, 7 October 2011 (UTC)

I think I have solved the issue with the particular banner now; if anyone is still having problems, please report them at m:Talk:Fundraising 2011 (for fundraising banners) or m:Talk:CentralNotice (for other banners, or issues regarding both fundraiser banners and other banners), where it is more likely to be seen by the relevant people. Jsoby (talk) 14:26, 7 October 2011 (UTC)

Does it still not work for you, John? It works for me, but if it's still broken for you I don't know what's wrong. :/ Jsoby (talk) 15:46, 7 October 2011 (UTC)
I assume it is something to do with the recent "upgrade". It would be great if it could be fixed, it is highly irritating. --John (talk) 15:54, 7 October 2011 (UTC)
Maybe, but my question still stands: Does it still not work for you? Because I thought I had fixed it. Jsoby (talk) 16:02, 7 October 2011 (UTC)
Hey, thank you, it seems to be fixed. Nice work. --John (talk) 16:47, 7 October 2011 (UTC)

They're baaack

I have the javascript/CSS thing below installed in addition to whatever fix was done here...but they're starting to pop up again. Not quite as often, but... - The Bushranger One ping only 19:44, 7 October 2011 (UTC)

Template:Infobox civil conflict

I don't think this templates way of doing images and pixel sizes is good. Can someone help? Thanks. - Peregrine Fisher (talk) 04:15, 7 October 2011 (UTC)

There's nothing in the template which mandates large image sizes. You've just misunderstood the documentation. See the talk page. Chris Cunningham (user:thumperward) - talk 09:22, 7 October 2011 (UTC)

Unwatching with Popups now has additional (unwelcome) prompt

I don't know if this is related to the recent upgrade or is just another odd issue with popups. Until recently, I could remove an article from my watchlist by hovering over it and selecting "unwatch" from the popups menu with one click that opened a page confirming the article had been removed. Now, that same menu option opens a page that requires me to confirm that I want to remove the article. What used to already be slightly annoying and took two clicks - one to unwatch the article and another to close the unnecessary new page confirming its removal - now takes three clicks. We've gone from bad to worse. Any ideas how to fix this or to whom I should direct my annoyance? ElKevbo (talk) 05:34, 7 October 2011 (UTC)

This is a result of bugzilla:27655 and bugzilla:29070. T. Canens (talk) 07:33, 7 October 2011 (UTC)
You should post at Wikipedia talk:Tools/Navigation popups to try to get the problem fixed. Anomie 10:59, 7 October 2011 (UTC)
I'm confused and I seem to be getting mixed messages. Is this a bug in the new software that is being fixed or is this an issue with popups that needs to be fixed locally? Or both? ElKevbo (talk) 14:22, 7 October 2011 (UTC)
Because of an extra security measure in 1.18, the popups functionality is no longer working, and popups will need to be fixed locally. —TheDJ (talkcontribs) 14:27, 7 October 2011 (UTC)
Ok. Thanks so much! ElKevbo (talk) 14:29, 7 October 2011 (UTC)

Sortable Tables glitch

Hi,

Since the 1.18 update, my sortable table on my user page has a few problems with secondary sort key. eg if you press age, then name, some are correct(eg Amanda Kimmel), but others arent (eg Ami Cusack). Any help would be appreciated. --Kangaross1989 (talk) 05:39, 7 October 2011 (UTC)

The behavior changed, because we use a new, more advanced tablesorter. This new system requires using the shift-key in order to do multiple column sort. bugzilla:31255. —TheDJ (talkcontribs) 12:08, 7 October 2011 (UTC)

Thumb images next to sortable table pushes the table underneath it

Thumb images next to sortable table pushes the table underneath it. and increase the inside of the table. Prior to the update it wasn't the case.

If you look at the exemple bellow you will see that until the page fully loads the text and images sit well next to the table but once the sorting images kicks in it pushes it and make the inside table bigger bcuz of the arrow images.

  – HonorTheKing (talk) 12:00, 7 October 2011 (UTC)

This depends on the width of your screen, the table will take up as much space as it requires. That might be slightly more that before, but you should never depend much on the width of an element. (Think also about mobile phones for instance). —TheDJ (talkcontribs) 12:16, 7 October 2011 (UTC)

References look diffrent in Firefox and IE

References in firefox have two refs in one line. while in IE only one ref in one line. I refer when it uses - {{reflist|colwidth=30em}} OR {{reflist|2}} Not talking about the normal {{reflist}} which post 1 ref per line in both browsers.
  – HonorTheKing (talk) 12:00, 7 October 2011 (UTC)

This is because many browsers (including IE) don't properly support columns yet. It also explains this in the documentation of Reflist. —TheDJ (talkcontribs) 12:06, 7 October 2011 (UTC)
"Many" browsers? As far as I was aware IE is the only modern browser which doesn't support CSS columns (at least using the -moz and -webkit implementation hacks, which we include in all of the relevant templates). Chris Cunningham (user:thumperward) - talk 13:57, 7 October 2011 (UTC)
Actually the chrome and safari implementations are a tad buggy as well (cutting of some lines in incorrect ways), and Opera ahs also had it's issues. Especially the colwidth variants of them. —TheDJ (talkcontribs) 14:07, 7 October 2011 (UTC)
IE10 supports columns; see WP:REFCOLS. Chrome and Safari appear to be mishandling the column-width selector, thus |n= works, but |colwidth= renders oddly. ---— Gadget850 (Ed) talk 16:54, 7 October 2011 (UTC)

connection difficulties

  Resolved
 – fixed
Extended content

I'm having problems connecting to Wikipedia using IE8. A box appears inviting me to report the problem to Microsoft, which I have done several times. If I click the "debug" button, IE often (not always) manages to display the page I want. This has been going on for several days. All the other sites I use work fine. I saw a bar a few days ago that said that maintenance work was being done that might cause connection difficulties. The bar has now gone, but the difficulties persist. Can this be fixed, please?! DOwenWilliams (talk) 15:06, 7 October 2011 (UTC)

It's been funny for a while. Probably the roll out of Media-wiki 1.18? If so the fix is to wait. Rich Farmbrough, 15:54, 7 October 2011 (UTC).
We just rolled out a fix for an 'IE8 crashes a lot' bug today, so check if this is now resolved for you. --brion (talk) 00:49, 8 October 2011 (UTC)
Yes. It works. Thanks. DOwenWilliams (talk) 03:00, 8 October 2011 (UTC)

I just spotted an odd bug with using piped links. If the pipe is immediately followed by the word "revision"

[http://en.wikipedia.org/w/index.php?title=List_of_plants_used_in_herbalism&oldid=452788457|revision at 01:14, 28 September 2011]
Gives at 01:14, 28 September 2011
[http://en.wikipedia.org/w/index.php?title=List_of_plants_used_in_herbalism&oldid=452788457|other word at 01:14, 28 September 2011]
Gives word at 01:14, 28 September 2011

in the first the link gets messed up and the word revision is not displayed.--Salix (talk): 22:34, 7 October 2011 (UTC)

Pipes aren't used in links using single brackets; the first space character is the delimiter instead. 28bytes (talk) 22:39, 7 October 2011 (UTC)

Sorting

Acknowledging that this function has now been corrupted, when can we expect a resolution? Are we going to be expected to re-code hundreds of articles as a result of this 'update'? The Rambling Man (talk) 19:14, 5 October 2011 (UTC)

Much of the problem is that the old code had no specification or tests at all so nobody could tell what was supposed to work or not or if any change would do anything good or bad. The new code has more test cases... but may also be yet incomplete.
The more specific things we know broke, the more we can either fix right off or replace back with the old code (with tests this time) until the new code works better. --brion (talk) 01:27, 6 October 2011 (UTC)
That's just gobsmacking. So much has been broken today that it's very obvious none of the developers have a clue about testing. Malleus Fatuorum 01:31, 6 October 2011 (UTC)
mw:Manual:JavaScript_unit_testing Your patches and bug reports are welcome! I'll certainly agree that we've had.... basically NO good testing for many years, but there've been big improvements this year so far. I wouldn't say it's ideal since, you know, it's not, so I assume that telling the truth is better than just pretending it's perfect? --brion (talk) 01:38, 6 October 2011 (UTC)
Honesty is certainly a welcome change, but surely the first rule of software development in the 21st century is to consider the tests before implementing the software. There is absolutely no excuse for today's debacle. Malleus Fatuorum 01:49, 6 October 2011 (UTC)
I can't go back in time, so I'll just hand you this thread. --brion (talk) 01:56, 6 October 2011 (UTC)

Note that I'd recommend we go back to the old sorting code until the new code gets more thoroughly resolved, but we still need to know things that broke so we can actually go test and fix them. Expect a flip back to the old sorting code tomorrow if nobody beats me to it. --brion (talk) 02:37, 6 October 2011 (UTC)

I think the suggestion being made is that you test first, see what breaks in a test environment, fix them, and then roll out. I would have thought this would be pretty much standard procedure, but I guess Wikipedia always has to be different...--Kotniski (talk) 07:06, 6 October 2011 (UTC)

Are there in fact remaining issues with sortable tables? Everything I can find cited on this page is listed as having now been fixed. If there are still other problems I'll go ahead and flip back the old code pending fixes; otherwise I suppose we'll leave it in place. --brion (talk) 18:57, 6 October 2011 (UTC)

Another issue is when clicking on the header cell of an unsortable column in a sortable table, the sort image of the last sorted column reverts to the neutral sort image, so you can no longer see in which direction and on which column the table is sorted. Pretty minor, but there it is. — Bility (talk) 22:08, 6 October 2011 (UTC)
Sortable tables still don't seem to be working. See, for example, List of federal judges appointed by Barack Obama#Courts of Appeals. The code says "sortable wikitable" and, until a few days ago, you could sort the columns without any difficulty. Now, the sort buttons don't even appear. --Lincolnite (talk) 05:18, 7 October 2011 (UTC)

References break sorting. See [19] and see how CU stats sort correctly in the September column. Then see [20] (desired revision) and notice how they don't. –xenotalk 17:23, 7 October 2011 (UTC)

You need to specify the attribute data-sort-type on table columns, when you are using mixed content in table cells. Unfortunately, it seems someone did not deploy the change, or perhaps it's just a setting of the servers, that actually allows "data-*" on wikimedia wikis. I have create bugzilla:31527 for this problem. The tablesorter should not have been deployed without this data attribute being allowed. —TheDJ (talkcontribs) 13:39, 8 October 2011 (UTC)

toolbar buttons not working on IE?

  Resolved
 – Now deployed —TheDJ (talkcontribs) 13:15, 8 October 2011 (UTC)

Reported, fixed, not yet deployed: toolbar buttons not workingMarkAHershberger 14:57, 6 October 2011 (UTC)

Extended content

The toolbar buttons (signature, markup, etc.) aren't working for me on IE. Is anyone else experiencing this? --Ixfd64 (talk) 04:54, 6 October 2011 (UTC)

toolbar buttons not working☠MarkAHershberger☣ (talk) 05:14, 6 October 2011 (UTC)
They don't work for me either and apparently don't work in IE 7 and 8 period (some else also tried). See the discussion above. Voceditenore (talk) 07:47, 6 October 2011 (UTC)
Same for me, using IE 8. Ruigeroeland (talk) 08:02, 6 October 2011 (UTC)

Losing my editing when I go to other pages

  Resolved
 – Duplicate of #Change in behavior of browser "Back" button

(sounds like the same issue) — MarkAHershberger 20:24, 6 October 2011 (UTC)

It may be resolved in the sense that we know it's a duplicate of another issue, but it is not resolved in the normal sense, i.e. that it is fixed. —Anomalocaris (talk) 20:55, 7 October 2011 (UTC)

Extended content

Argh! For years, I would edit a page, click "Preview", and click a link in the text I just typed to make sure it didn't go to a disambig page or the wrong page; or would research a little more on the topic of the link; then I would click my browser's "Back" button until I was back on the editing page. This all worked until today! My editing changes are lost when I get back to the editing page, as though I had *just* clicked the "edit" link in the section header and had not done any typing yet. What changed? Comet Tuttle (talk) 18:43, 6 October 2011 (UTC)

Thanks, I was about to report this same issue! The crazy thing is, it's intermittent. I just edited Parentheses pipe shortcut doesn't work in <ref> above, and went to Show preview and back many times and never lost anything, but in editing other articles, I've found myself in Comet Tuttle's situation several times in the past day or two. —Anomalocaris (talk) 18:53, 6 October 2011 (UTC)
Are either of you running twinkle? Because since I turned it off, I seem to be able to edit as normal. DrKiernan (talk) 19:00, 6 October 2011 (UTC)
I'm not. Here it's Firefox 3.6.23 under Windows Vista with AdblockPlus and various other plugins that I don't think will modify the page being viewed. I'm set to Monobook, old school style. Comet Tuttle (talk) 19:11, 6 October 2011 (UTC)
I don't think so, but what is twinkle and how do I know if I'm running it? —Anomalocaris (talk) 19:21, 6 October 2011 (UTC)
WP:TWINKLE. Check the the "Twinkle" gadget in the Gadgets section of your Preferences page. --Tagishsimon (talk) 19:35, 6 October 2011 (UTC)
I don't use Twinkle; in fact I don't use any Gadgets. —Anomalocaris (talk) 20:14, 6 October 2011 (UTC)
Oh dear, it was going so well but now it's just happened again. DrKiernan (talk) 19:46, 6 October 2011 (UTC)
See the comment Problem with Internet Explorer I wrote earlier today. The bug is identified and has still to be solved. --Réginald alias Meneerke bloem (To reply) 19:26, 6 October 2011 (UTC)
Réginald alias Meneerke bloe, your Bugzilla note says "When opening a page on the English wikipedia with Internet Explorer, an error message appears and the page has to be re-opened. When making changes and visualising them, the changes are lost, so that I had to make the changes again and to save them without visualising them." I'm sorry for your difficulties, but my problem is different. I'm using Firefox, and I don't get any error message. My changes just disappear, both in the main editing window and in the edit summary. —Anomalocaris (talk) 20:00, 6 October 2011 (UTC)
I don't get a problem, but I open the link with right click and Open in new window, then before I read it, I click Back on the first window. I'm just naturally suspicious... This is in Firefox. Peridon (talk) 16:38, 8 October 2011 (UTC)

preloaded deletion reasons

  Resolved
 – duplicate of #Old block reason doesn't pre-load

MarkAHershberger 19:43, 6 October 2011 (UTC)

Extended content

Normally, when deleting a CSD-tagged page the reason in the tag is automatically preloaded when one goes to delete the page. Doesn't seem to be happening today. Beeblebrox (talk) 19:30, 6 October 2011 (UTC)

Hardly extended content, but whatever. Is it going to be fixed or what? Also, did the new software solve any problems, or is it specialized to only cause them? Beeblebrox (talk) 23:17, 6 October 2011 (UTC)
The biggest improvements in MW1.18 are for wikis in languages other than english. You can judge for yourself whether any of the other changes are of interest to you; but I'd suggest that [21], [2] and [3] are probably obvious candidates to be qualified as 'improvements'. Happymelon 10:08, 7 October 2011 (UTC)
I can't really make any sense out of those three links, so I have no idea whether they represent improvements or not, but I'm willing to take your word for it. Was just venting a bit as there seem to be a lot issues springing up that apparently were not anticipated. Beeblebrox (talk) 19:15, 7 October 2011 (UTC)
Check out what's new for a more readable list of improvements. — MarkAHershberger 22:34, 7 October 2011 (UTC)
I've been getting this - and it coincides with times when my compressing of 'discussion' to 'talk' doesn't come in (it's some .js thing - no idea how it works), and the Twinkle buttons don't arrive either. I've been reloading the window to get things back - sometimes with Ctrl-Reload if it doesn't work the first two times (usually does). Damned irritating. Peridon (talk) 16:44, 8 October 2011 (UTC)

Fundraiser

This latest box-at-the-top is really annoying me. I've read it, I'm not sharing my story, and I want it to go away. Usually there's a 'hide' button on these things at the top so you can get rid of notices you've seen and don't want any more. This damn thing can be closed, but it reappears in the very next window. It's not a life threatening problem - just irritating. This probably isn't the right place, but I don't know of another off hand. Hopefully the person that set it up will hear about this and alter it so it can be tidied away by people who don't like being nagged. Peridon (talk) 12:34, 7 October 2011 (UTC)

I agree. Also, the "suppress fundraiser banners" option in the preferences seems gone. Regards SoWhy 12:49, 7 October 2011 (UTC)
I also want these annoying messages to go away. Any help on how to achieve this would be appreciated. Toshio Yamaguchi (talk) 12:53, 7 October 2011 (UTC)

I click hide and then 2 pages later it will come up again or another banner will come up about mobile phone. What the hell is wrong with them? I looked in my preferences for the supress banners and its not there. ♦ Dr. Blofeld 12:57, 7 October 2011 (UTC)

Can someone copy paste and list here the exact text of any banner that is doing that ? I'm not seeing them myself, perhaps they are geotargeted or something. —TheDJ (talkcontribs) 13:15, 7 October 2011 (UTC)
Adding the following to special:Mypage/skin.css should hide it -
#centralNotice  .siteNoticePic {
    display: none;
}
There's a bunch of other stuff you can hide, check User:Xeno/monobook.css. Of course you will need to accept whatever risks may come along with hiding central, site, watchlist, and edit notices. –xenotalk 13:30, 7 October 2011 (UTC)
I do want to see notices - but once only. A hide button is all I ask. I know this is a WikiMedia thing (I tried opening it, but it doesn't seem to accept that as proof that I've read it...), but is it possible for someone to tell them how annoying (and counter-productive) it is? Peridon (talk) 13:54, 7 October 2011 (UTC)

I think I have solved the issue with the particular banner now; if anyone is still having problems, please report them at m:Talk:Fundraising 2011 (for fundraising banners) or m:Talk:CentralNotice (for other banners, or issues regarding both fundraiser banners and other banners), where it is more likely to be seen by the relevant people. Jsoby (talk) 14:26, 7 October 2011 (UTC)

I am still having problems, and have posted at meta about it. Peridon (talk) 20:20, 7 October 2011 (UTC)
Don't know what they've done, but the banner hasn't appeared today. Peridon (talk) 15:28, 8 October 2011 (UTC)

Persisting bug with IE

The persisting bug with IE precludes the "Show preview" of modifications/corrections. This involves that errors/typos made during these modifications/corrections are not detected before saving and that these possible errors/typos have to be corrected in a second step.
Please solve this issue ASAP. --Réginald alias Meneerke bloem (To reply) 10:35, 7 October 2011 (UTC)

I am also encountering this one, along with the "crashes every two or three pages" that everybody has who's stuck with IE8. --Orangemike (talk) 19:07, 7 October 2011 (UTC)
Is this fixed now that IE8 (should be) working? — MarkAHershberger(talk) 05:02, 9 October 2011 (UTC)

Script not working on Watchlist

Using Firefox. Opening my watchlist I get a message to say that a script on the page has stopped working - http://bits.wikimedia.org/en.wikipedia.org/load.php?debug=false&lang=en&modules=jquery%2Cmediawiki&only=scripts&skin=monobook&version=20111005T084642Z:99  pablo 15:51, 7 October 2011 (UTC)

This is another "bits.wikimedia.org" problem similar to the one I reported here a few days back. I'm still not sure why the browser spends so much time "waiting for bits.mediawiki.org". Rich Farmbrough, 15:57, 7 October 2011 (UTC).
I'm reporting this as as something to keep track of, but I'd like more reports before I take it further. — MarkAHershberger 20:44, 8 October 2011 (UTC)

Tab underlining

  Resolved
 – fixed

Could a tech-savvy user tell me please how to stop the WP tabs from being underlined? "Project page"/"Article", "Discussion", "Read", "Edit" and "View history" are all underlined now when they weren't before. This is not the case when logged-out, but when I'm logged in, I have the preference for wiki-links to be underlined... It's just that I don't want the tabs underlined at the top of each page – and I'm not familiar with CSS in order to help myself. Jared Preston (talk) 18:47, 7 October 2011 (UTC)

If you're using Vector, the following CSS code will get rid of the underlines:
#mw-head a {
  text-decoration: none !important;
}
Edokter (talk) — 21:51, 7 October 2011 (UTC)
If they were not underlined before, then I would consider that to be a bug and the behavior now fixed. The preference talks about underlining ALL links, not just wikilinks. —TheDJ (talkcontribs) 12:51, 8 October 2011 (UTC)

Category Navigation

I've got an issue with navigating through a large category using the next & prev links.

If I view any large category which has a category toc (for example: Category:Hidden categories but any with a table of contents will do) and then select a letter from the table of contents. When I then try to use the previous 200 link, it does not work. -- WOSlinker (talk) 15:15, 8 October 2011 (UTC)

I can confirm that. Right now I'm using Firefox. Chris857 (talk) 19:22, 8 October 2011 (UTC)
This sounds like a recurrence of a problem which appeared soon after the MW1.17 deployment (see Wikipedia:Village pump (technical)/Archive 87#categorically random categories) - the URLs are constructed incorrectly, and it concerns whether &from= or &pagefrom= should be used in the URL. The workaround is to click the letter in the category TOC, then go to your browser's address bar and alter the &from= to a &pagefrom= and press Return. The "previous 200" link should then work until you next select a letter in the category TOC. --Redrose64 (talk) 19:23, 8 October 2011 (UTC)
You have to use &subcatfrom= when navigating Category:Hidden categories though. Is the a bug enter for this issue in Bugzilla? -- WOSlinker (talk) 20:25, 8 October 2011 (UTC)
See Category Navigation brokenMarkAHershberger(talk) 03:52, 9 October 2011 (UTC)

Not all the tools display

Not sure if this is the right place to report this, but I've noticed since the recent upgrade that not all the tools are displayed whenever a new article is created or some articles are edited. This is particularly true with the redirect and ref tools. I thought at first it might be because I'd recently changed my username, but this doesn't appear to be the case as it works with some long-established articles. Apologies if this has already been reported. Cheers Paul MacDermott (talk) 19:02, 8 October 2011 (UTC)

Maybe the same as this problem on ptwiki? — MarkAHershberger 21:01, 8 October 2011 (UTC)
Forgot to reply to this, so apologies. Looks similar. Still experiencing the issue, so hopefully a fix will be available soon. Paul MacDermott (talk) 12:46, 14 October 2011 (UTC)

Global preferences?

Is there anyway to set my preferences globally for all Wikimedia projects? I much prefer Monobook to Vector, and find it a bit annoying that if I click over to another wiki I have to set the preference yet again. Can I set this in one place and have it cver all my MediaWiki accounts? LadyofShalott 23:57, 8 October 2011 (UTC)

No. T16950 ---— Gadget850 (Ed) talk 00:23, 9 October 2011 (UTC)
Drat. Thanks, Ed. LadyofShalott 00:36, 9 October 2011 (UTC)

Odd behaviour on "User Contributions" on previously blocked users.

Added this as a comment to Bugzilla:31403. — MarkAHershberger(talk) 03:33, 9 October 2011 (UTC)

This link http://en.wikipedia.org/wiki/Special:Contributions/71.125.52.143 is telling me the user is blocked - where one can clearly see that it well pass one week after the 11th September. This page http://en.wikipedia.org/wiki/Special:Block/71.125.52.143 is fine and says "This user has been blocked previously".  Ronhjones  (Talk) 01:19, 9 October 2011 (UTC)

FYI: before MW 1.18 it was like this: if IP was blocked indirectly (through range block or autoblock) then on IP contribs page:
  1. sysops would see "change block" link instead of "block" under heading;
  2. if IP was ever blocked directly MediaWiki would display display that outdated block message.
This behaviour seems to be fixed for range blocks (I really miss #1 though) but maybe not for autoblocks?
AlexSm 04:06, 9 October 2011 (UTC)
Perhaps this is another case of bugzilla:27858 (presuming we were totally wrong about timezones having anything to do with that issue). Bawolff (talk) 04:12, 9 October 2011 (UTC)

Reftool 1.0 failing when filling fields from URL

  Resolved
 – TheDJ (talkcontribs) 21:46, 9 October 2011 (UTC)

Reported: webkit problems with reftool Local mod, report errors here: Wikipedia_talk:RefToolbar

I use Reftools 1.0 because it can grab data and fill in fields when given the source URL. When I use it today, instead of filling in the fields, I find that I'm no longer in edit more and I'm looking at the "Read" version of the article. The browser is Chrome-- Whpq (talk) 15:28, 5 October 2011 (UTC)

I have the same problem, not just in Chrome but also in Safari. --joe deckertalk to me 13:59, 6 October 2011 (UTC)
I have the same problem with Firefox 7.0.1. -- Alan Liefting (talk) - 07:49, 8 October 2011 (UTC)
Got it. Apparently, an image control acts as a submit button. I think there was a bug in older versions of jquery we used before 1.18 deploy, that was canceling this behavior. After the 1.18 upgrade, the issue in jquery was fixed, and the issue in the script became visible. Now fixed. Might take a while to propagate to all users. —TheDJ (talkcontribs) 21:46, 9 October 2011 (UTC)
Wacky little bug there, fascinating. Thanks so much for digging into this! --joe deckertalk to me 21:52, 9 October 2011 (UTC)
The fix has not propagated thought to me yet but while I have your attention is the ISBN autofill working? It has not worked for me for quite some time - as in prior to MW1.18. -- Alan Liefting (talk) - 04:58, 10 October 2011 (UTC)

Wikipedia Vulnerable for Attack

FYI I guess you know but, anyway : [removed per WP:BEANS] 87.64.24.15 (talk) 11:38, 6 October 2011 (UTC)

This kind of thing should be reported to security wikimedia.org. I'm emailing them now. MER-C 11:46, 6 October 2011 (UTC)
Done. Thanks, but please avoid posting such information in public fora before the problem(s) are fixed. MER-C 11:54, 6 October 2011 (UTC)
Ok, sorry... 87.64.168.237 (talk) 16:22, 9 October 2011 (UTC)

markBlocked script not working

This script User:NuclearWarfare/Mark-blocked script.js which nicely showed all blocked users names strikedthrough, is now not working.  Ronhjones  (Talk) 01:12, 9 October 2011 (UTC)

Added a comment to User_talk:NuclearWarfare pointing to this problem. — MarkAHershberger(talk) 03:47, 9 October 2011 (UTC)
I just import that script from ruwiki; I would have no idea how to fix it. I don't know who might be able to help, perhaps Anomie or Xeno? NW (Talk) 03:49, 9 October 2011 (UTC)
The script was broken in MediaWiki 1.18 and I happened to be the person who fixed it in ruwiki on Sep 29. Unfortunately just a few hours before that you switched to your local (still broken) version. Next time please do this: 1) try once again the script from the orignial source and 2) try to contact the author / maintainer directly. — AlexSm 04:26, 9 October 2011 (UTC)
Thanks Alex - I was trying both versions at the time and neither (at that time) worked - I will sort out the local version.  Ronhjones  (Talk) 20:57, 9 October 2011 (UTC)

Technical problem message on invalid page

I accidentally entered a link of the form http://en.wikipedia.org/wiki/User:/xxx. It is obviously invalid, but I would not have expected to get "Our servers are currently experiencing a technical problem." ---— Gadget850 (Ed) talk 13:17, 9 October 2011 (UTC)

"PHP fatal error in /usr/local/apache/common-local/php-1.18/extensions/WikiLove/WikiLove.hooks.php line 185: Call to a member function getTalkPage() on a non-object". Looks like WikiLove messed up. I'll put it on Bugzilla. Ucucha (talk) 13:46, 9 October 2011 (UTC)
If I turn off the love, I get the more expected "User account "" is not registered." ---— Gadget850 (Ed) talk 13:53, 9 October 2011 (UTC)
Yes; the reason is that WikiLove doesn't error-check when it calls Title::newFromText(), so it tries to treat NULL as an object, which generates a PHP fatal error. I've reported it in bugzilla:31557. Ucucha (talk) 14:00, 9 October 2011 (UTC)

Cite news template "accessdate" parameter not working

The "accessdate" parameter on the cite news template is not working. Example: "Twenty Dead in Havana" reference in 1906 Florida hurricane; there is an |accessdate= parameter but it isn't working. HurricaneFan25 14:42, 9 October 2011 (UTC)

Without even editing, I can tell you don't have |url=, which means |accessdate= will not show. ---— Gadget850 (Ed) talk 14:47, 9 October 2011 (UTC)
Confirmed (but I had to edit) — MarkAHershberger(talk) 15:08, 9 October 2011 (UTC)
It is abnormal to give an access date for a static version of a newspaper, such as a printed or microfilm copy. But in general, an access date could be given for media that gets updated, even though that media does not have a URL. For example, help files included with the Microsoft operating system would be updated periodically, so it could be useful to specify an access date, even though there is no URL. Of course, that particular example would not use the "Cite news" template. Jc3s5h (talk) 15:31, 9 October 2011 (UTC)
The clue here is that the title is not linked, thus there is no |url=. Now documented at WP:CS1PROBS. ---— Gadget850 (Ed) talk 16:46, 9 October 2011 (UTC)

[edit] in some articles not shown in some cases

Before someone ask, yes I can edit the article from the top edit botton.

now the [edit] in the aricle body sections do not longer show in some pages. See here - Jersey Shore (TV series), it doesn't show. While in Manchester United F.C. I do see the [Edit] in each section. Just note that im able to edit both articles using the Edit in the top between Read and View history. but the edit in each section is missing and neccesery in the first article.
  – HonorTheKing (talk) 16:55, 9 October 2011 (UTC)

Probably a __NOEDITSECTION__ magic word in the article or a template. Looking now. ---— Gadget850 (Ed) talk 17:40, 9 October 2011 (UTC)
Now it shown. Guess someone changed something.
  – HonorTheKing (talk) 18:01, 9 October 2011 (UTC)
Someone just edited a table. Perhaps something stuck in cache and the change purged it. ---— Gadget850 (Ed) talk 18:02, 9 October 2011 (UTC)
See #No section edit links above. Known issue. Edokter (talk) — 18:14, 9 October 2011 (UTC)

1.18 issues tagged as resolved

bug in #tag:ref parser function

  Resolved
 – Should be working now. Purge pages if error is still appearing. mc10 (t/c) 05:42, 6 October 2011 (UTC)
Extended content

I guess this is yet another problem with the MW 1.18 rollout above.

When a <ref> tag is used within a note which is itself within a #tag:ref function, the superscript appears as a very long (and ugly) string (representing the internal link to the cite) instead of the normal small numeral. The link is still clickable and works fine, apart from the very ugly and off-putting display. Example at User:NSH001/sandbox.

--NSH001 (talk) 09:02, 5 October 2011 (UTC)

Confirmed. This breaks {{refn}}. I think we have seen this one before. ---— Gadget850 (Ed) talk 09:06, 5 October 2011 (UTC)
Err, does that mean it's going to be fixed? It's been working fine (for months, at least) until now. --NSH001 (talk) 09:12, 5 October 2011 (UTC)

See the section above this one, and see an example of the damage: Prince George of Denmark#Notes. DrKiernan (talk) 09:11, 5 October 2011 (UTC)

All refs on that page contain nested <ref> tags; I don't think that is even supported... Edokter (talk) — 09:17, 5 October 2011 (UTC)
Support was added in rev:33066. {{refn}} has been working until now. ---— Gadget850 (Ed) talk 09:29, 5 October 2011 (UTC)
It's the way suggested at WP:REFNEST when one wants to place a reference in a footnote. DrKiernan (talk) 09:25, 5 October 2011 (UTC)
For more examples, see Anna Essinger#Footnotes and Bunce Court School#Footnotes and Sir Arthur Harris, 1st Baronet#References. Hope this is fixed soon! Marrante (talk) 09:16, 5 October 2011 (UTC)
Already reported as T27417. ---— Gadget850 (Ed) talk 09:29, 5 October 2011 (UTC)

Has this been fixed? I'm not seeing it. Ucucha (talk) 11:33, 5 October 2011 (UTC)

Nope, still not working. --NSH001 (talk) 11:38, 5 October 2011 (UTC)

Per Happy-melon on previous issues:

The "uniq...quinu" strings are strip markers: placeholders that the parser inserts to say "more complicated content will go here". Exposing strip markers, which is what is happening here, occurs when badly-formed code in MediaWiki causes the parser to be reset, and lose its memory of which strip marker corresponds to which piece of special content.

Logged as T33374. If you see the strip marker below, then it is still broken; if you see "Fixed now!", then it is fixed. ---— Gadget850 (Ed) talk 12:07, 5 October 2011 (UTC) [note 1]

  1. ^ [1]
  1. ^ Fixed now!

I just visited Tropical Storm Debra (1978), one of my articles, and the "footnotes" section says "Tornadoes were confirmed in Texas,?UNIQ4f238ebe6a8c71c3-nowiki-00000031-QINU?4?UNIQ4f238ebe6a8c71c3-nowiki-00000032-QINU??UNIQ4f238ebe6a8c71c3-nowiki-00000034-QINU?5?UNIQ4f238ebe6a8c71c3-nowiki-00000035-QINU??UNIQ4f238ebe6a8c71c3-nowiki-00000037-QINU?6?UNIQ4f238ebe6a8c71c3-nowiki-00000038-QINU? Louisiana,?UNIQ4f238ebe6a8c71c3-nowiki-0000003A-QINU?4?UNIQ4f238ebe6a8c71c3-nowiki-0000003B-QINU??UNIQ4f238ebe6a8c71c3-nowiki-0000003D-QINU?7?UNIQ4f238ebe6a8c71c3-nowiki-0000003E-QINU? Mississippi,?UNIQ4f238ebe6a8c71c3-nowiki-00000040-QINU?4?UNIQ4f238ebe6a8c71c3-nowiki-00000041-QINU??UNIQ4f238ebe6a8c71c3-nowiki-00000043-QINU?8?UNIQ4f238ebe6a8c71c3-nowiki-00000044-QINU? Arkansas,?UNIQ4f238ebe6a8c71c3-nowiki-00000046-QINU?9?UNIQ4f238ebe6a8c71c3-nowiki-00000047-QINU? and Tennessee.?UNIQ4f238ebe6a8c71c3-nowiki-00000049-QINU?10?UNIQ4f238ebe6a8c71c3-nowiki-0000004A-QINU?" HurricaneFan25 13:16, 5 October 2011 (UTC)

Already reported; see #bug_in_.23tag:ref_parser_function. –Drilnoth (T/C) 13:17, 5 October 2011 (UTC)
See #bug in #tag:ref parser function above. ---— Gadget850 (Ed) talk 13:17, 5 October 2011 (UTC)
We have plenty of examples now. If you check progress on the Bugzilla page, you will see the the issue has been characterized and is being worked on. When it is fixed, the sample I provided above will start working. ---— Gadget850 (Ed) talk 20:00, 5 October 2011 (UTC)

  Fixed You may need to purge a page. Please update if you see more issues. ---— Gadget850 (Ed) talk 01:32, 6 October 2011 (UTC)

Had to add "?action=purge" to the URL, but after purging - hurrah, eet werked! Nice work devs. :) - The Bushranger One ping only 01:58, 6 October 2011 (UTC)
  Resolved
 – fixed
Extended content

Can't click on wikilinks in diffs (or, more usefully, get them to show in popups). A big time-waster. --NSH001 (talk) 10:13, 5 October 2011 (UTC)

That wasn't core functionality, that is a gadget which is currently broken as to my understanding. Peachey88 (T · C) 11:43, 5 October 2011 (UTC)
I think that should be reported on User_talk:Cacycle/wikEd. Helder 15:59, 5 October 2011 (UTC)


Watchlist doesn't collapse

  Resolved
Extended content
It says
Error: $that.attr("id") is undefined
Source File: https://secure.wikimedia.org/wikipedia/en/w/index.php?title=MediaWiki:JQuery-makeCollapsible.js&action=raw&ctype=text/javascript
Line: 230

on Special:Watchlist, and for pages where there have been multiple edits, those do not collapse into a single line per page. Ucucha (talk) 11:26, 5 October 2011 (UTC)

(edit conflict) I was just about to say that enhanced recent changes isn't working for me either, on recent changes or my watchlist. Grouping works, not collapsing. →Dynamic|cimanyD← (contact me) 11:30, 5 October 2011 (UTC)
To clarify, this is in Firefox 6.0.2 on Vector. The error console gives the error I pasted above twice, plus the following one:
Error: uncaught exception: Syntax error, unrecognized expression: [href*=&days=7]

I haven't tried turning off other scripts yet. Ucucha (talk) 11:43, 5 October 2011 (UTC)

This seems to be fixed after our local copy of the jQuery plugin "makeCollapsible" was removed. Could you confirm if your Special:Watchlist is working again? (Remember to clear your cache) Helder 14:34, 5 October 2011 (UTC)
It is, thanks. Ucucha (talk) 16:02, 5 October 2011 (UTC)
Working! →Dynamic|cimanyD← (contact me) 23:13, 5 October 2011 (UTC)

Classic skin sidebar

Reported: Classic skin has extra spacing in IE8MarkAHershberger 17:12, 6 October 2011 (UTC)

Extended content

(e/c/) I don't know if it's due to 1.18, or someones been altering the css/js, but I'm getting two separator bars between "Contact Wikipedia" and "Edit this page" on the sidebar of Classic Skin, when looking at editable pages. Not a major issue, but it looks wrong.—An  optimist on the run! 11:36, 5 October 2011 (UTC)

I'm seeing large gaps between the various sections, i.e. between Donate to Wikipedia and Search etc.... The Rambling Man (talk) 12:09, 5 October 2011 (UTC)
Can you provide a screenshot? Sumanah (talk) 22:41, 5 October 2011 (UTC)
I was seeing that too. I changed to the Vector skin, and it's OK now. Voceditenore (talk) 12:53, 5 October 2011 (UTC)
One shouldn't have to change to Vector for the sidebar to appear correctly. Some of us like MonoBook, and it's screwed up there as well. Deor (talk) 18:24, 5 October 2011 (UTC)
I have switched multiple times to both skins. I never see this problem. What browsers are the users using ? Perhaps it's something browser specific. —TheDJ (talkcontribs) 15:48, 6 October 2011 (UTC)
IE8 and Firefox 7 - I get the same problem in both.—An  optimist on the run! 15:50, 6 October 2011 (UTC)
See this image, which also shows the problems caused by the category spacing.—An  optimist on the run! 16:52, 6 October 2011 (UTC)
I'm also having this problem. Here's a screenshot. ~Asarlaí 05:43, 10 October 2011 (UTC)

Collapsible navboxes

  Resolved
 – Looks fine to me now. DrKiernan (talk) 08:07, 6 October 2011 (UTC)
Extended content

Navboxes no longer collapse. DrKiernan (talk) 12:01, 5 October 2011 (UTC)

Works for me. Can you give some examples? --NSH001 (talk) 12:22, 5 October 2011 (UTC)
All the boxes at the bottom of Prince Philip, Duke of Edinburgh used to collapse. Now they are all fully expanded for me. DrKiernan (talk) 12:43, 5 October 2011 (UTC)
Ah, they work in Explorer but not Firefox. (I've just switched to Firefox because I can't edit in Explorer anymore.) DrKiernan (talk) 12:47, 5 October 2011 (UTC)
They are all expanded for me in IE8. Problems were also reported at WP:HD regarding Steve Bruce. - David Biddulph (talk) 12:50, 5 October 2011 (UTC)
Works for me (Firefox 7.0.1, Ubuntu Linux 11.04). Off topic, that article has way too many categories. Takes up a whole screen on my browser since the apparent font size change. –Drilnoth (T/C) 13:17, 5 October 2011 (UTC)
WFM too (Firefox 9.0a2 on Win7 x64). I agree that the change in the categories font is pretty annoying though. Should be changed back imho. Regards SoWhy 13:36, 5 October 2011 (UTC)
If you think that's too many categories, you should look at his wife's article. And there are even more categories which are hidden with <!-- code. DrKiernan (talk) 13:37, 5 October 2011 (UTC)
Should be fixed after this changeTheDJ (talkcontribs) 16:01, 5 October 2011 (UTC)
Nope, they still don't collapse for me. Interestingly the problem only occurs when I am logged in, also I ticked 'Exclude me from feature experiments' and then 'Don't show the Article feedback widget on pages'. Both times this solved the problem very briefly, but then it came back. Just this second I changed my skin, and refreshed and the problem was fixed. I refreshed again and the problem returned.--EchetusXe 17:24, 5 October 2011 (UTC)
Navboxes seem to be collapsing normally for me. - The Bushranger One ping only 23:22, 5 October 2011 (UTC)
They look OK to me now too. I shall tag this as resolved for now, but just remove the tag if others still have problems. DrKiernan (talk) 08:07, 6 October 2011 (UTC)
Yeah, looks to be fixed now.--EchetusXe 08:57, 6 October 2011 (UTC)

The new Sorting breaks all tables arround Wikipedia which include two headers

  Resolved
 – fixed
Extended content

The issue is that the new Sorting applying in the past few days breaks all tables arround Wikipedia which includes two headers once you sort one.

Exemples (there are many more but posted the ones I use to edit):

  – HonorTheKing (talk) 12:00, 5 October 2011 (UTC)

Fixed in r99092. --Catrope (talk) 13:39, 6 October 2011 (UTC)


I can't see any sortable tables on either of those articles (using IE8). - David Biddulph (talk) 12:17, 5 October 2011 (UTC)
And looking at the wikisource it says unsortable. Where are you finding something sortable? - David Biddulph (talk) 12:19, 5 October 2011 (UTC)
Both tables are sortable, (IE9, Chrome, Firefox) not sure why you can't see them. exept of Ref they both sort. they sort for me but they break and merge the headers.
In addition, if you press the wikilinked in the header it sort the table and not go to the pressed linked.
Exemple: If you press FA Cup it sort it but not direct to it. It used to direct while pressing the names, and only by pressing the up and down icons it sorted.
  – HonorTheKing (talk) 12:32, 5 October 2011 (UTC)
The second table titled "List of Manchester United F.C. players with at least 100 appearances". In IE9 and FF7, clicking on a sort arrow creates a second header row with the Appearances title cell in the leftmost column followed by the other title cells. In IE9, when I clock on a sort arrow, it pushes the table below the images. ---— Gadget850 (Ed) talk 12:37, 5 October 2011 (UTC)
Looking again at the wikisource the one in List of Manchester United F.C. seasons does say sortable and one of the tables in List of Manchester United F.C. players does too, though others are unsortable. I don't know why they don't appear sortable for me. Perhaps it's connected with the new software? - David Biddulph (talk) 12:44, 5 October 2011 (UTC)
The sort buttons are still there, but are now smaller (and nicer, I think). As far as I can remember, though, sorting tables with more than one row of headers has always been dodgy, which is why, when I create a new sortable table, I restrict it to one row of headers. --NSH001 (talk) 12:53, 5 October 2011 (UTC)
They sorted well before. No issues or something like that. The few issues were fixed few weeks ago after improving all tables after FLC talk for ACCESS. In past and now aswell, the two headers always included the sorting icons in the first header while the second didn't have one, but it was able to sort.
The Russian Wikipedia uses the same table format and as it uses the old Wikipedia Version you can see how it works there. Russian Page for ManUtd players
  – HonorTheKing (talk) 12:59, 5 October 2011 (UTC)
Also see List of India women ODI cricketers. Has the sort problem, hidden sort arrows and also the "unsortable rows" bit doesn't work -- it sends the unsortables to the top. —SpacemanSpiff 13:39, 5 October 2011 (UTC)
If the sort buttons are there but smaller, for me they are so small as to be invisible and so as to be ineffective. I still cannot sort those tables. The Russian one will sort, the English version won't. - David Biddulph (talk) 14:51, 5 October 2011 (UTC)
The fact that this (multispan headers with sorting) worked before, was a side effect of the sortable script, not being able to properly handle rowcolspans. As such the fact that it worked like this a few weeks ago, is by all intents and purposes an accident. Now colspan and rowspan functionality actually works (one of the most longstanding requests in bugzilla in think). Now this breaks. I say, file as a bugreport, and when someone has a lot of time to spare, they can dive into the tablesorter code :D —TheDJ (talkcontribs) 15:58, 5 October 2011 (UTC)
I have similar problem with sorting on 2010–11 HNK Hajduk Split season. Using Firefox 7.0.1., it breaks Appearances and goals table when trying to sort it and the table "freezes". Previously, I could sort by number, position, player name and total appearances. Now those sort arrows are gone and only sort arrows are that under apps and goals. While in IE8 and Opera 11.51, sort arrows aren't visible at all. Dr. Vicodine (talk) 18:50, 5 October 2011 (UTC)

Looks like this is fixed, but will probably require purging the page. Lets get some feedback on this. ---— Gadget850 (Ed) talk 01:54, 6 October 2011 (UTC)

Iv'e purged the page and cleaned my browser and the sorting problem still there. Only thing fixed is that now you can click the wikilinked in the header as it fixed few bellow.
  – HonorTheKing (talk) 08:45, 6 October 2011 (UTC)
The problem with the wikilinks is solved yes, the other issue is bugzilla:31420. —TheDJ (talkcontribs) 09:19, 6 October 2011 (UTC)

Namespace option on contribution history

  Resolved
 – fixed

I'll be short. My User contributions is normal and functional otherwise, but it's missing the droplist for Search by namespace. It works normally in all other wikipedias and Commons too, but not here. I want it back :( Pitke (talk) 06:35, 5 October 2011 (UTC)

This function was removed at the request of Domas, one of the developers. No-one knows what his reason for requesting this change was. Apparently, "if we don't hear from him [Domas] by the end of this week, then I'll suggest that the change be backed out." See mediazilla:31197. — This, that, and the other (talk) 06:43, 5 October 2011 (UTC)
Change has been reverted, see Bugzilla. --Catrope (talk) 13:52, 6 October 2011 (UTC)
Extended content
I want it back as well. I often used that feature. -- Alan Liefting (talk) - 07:42, 5 October 2011 (UTC)
I also use the function, though if there were a very good reason for removing it I suppose I could live without it.   Will Beback  talk  07:54, 5 October 2011 (UTC)
Yes, let's have that back. –xenotalk 12:38, 5 October 2011 (UTC)
Agreed - I want the feature back. I use it to track my deletion discussions where I don't want to add them to my watchlist. SchuminWeb (Talk) 03:43, 6 October 2011 (UTC)
The entries in my Watchlist are now all expanded rather than being collapsed. Would this be a related issue? -- Alan Liefting (talk) - 07:50, 5 October 2011 (UTC)
Not really related. See this other section. It seems this edit fixed it. Helder 15:23, 5 October 2011 (UTC)

For me, the contribs dropdown menu for search by namespace is gone both here and on Wikimedia Commons. – Athaenara 07:58, 5 October 2011 (UTC)

I have both of those probs   Sp33dyphil "Ad astra" 08:43, 5 October 2011 (UTC)
I also hope the namespace option here can be returned; I found it very useful in keeping tracking of discussions outside of articlespace. Hullaballoo Wolfowitz (talk) 16:36, 5 October 2011 (UTC)
This is an extremely useful feature; I hope it is restored soon. For one thing, my bot does a lot of logging to a subpage in userspace and it is very handy to filter those contributions out and just view the articlespace changes it's made. 28bytes (talk) 17:39, 5 October 2011 (UTC)
It appears to have been removed intentionally. :( –Drilnoth (T/C) 18:00, 5 October 2011 (UTC)
Then it should be restored as an option. It's not just useful, it's necessary. - The Bushranger One ping only 19:30, 5 October 2011 (UTC)
Agreed, it should be restored. --NSH001 (talk) 21:57, 5 October 2011 (UTC)
A very useful filtering option; should be restored. . . Mean as custard (talk) 07:53, 6 October 2011 (UTC)

There used to be (12 hours ago or so) a "Namespace" option on the contributions history pages. Did it disappear for everyone? I've tried switching my interface language back to English from Dutch, and my skin from Monobook back to Vector, and it stays gone.—Kww(talk) 11:34, 5 October 2011 (UTC)

Gone for me, missed your while I was writing mine below.--Crossmr (talk) 11:42, 5 October 2011 (UTC)

Did I miss a memo? When did this happen? I was just checking out contribs and I can no longer select a drop-box to filter by name space (article/talk/wikipedia/etc). This was insanely useful. Why was this removed?--Crossmr (talk) 11:42, 5 October 2011 (UTC)

Confirmed. This did work yesterday. ---— Gadget850 (Ed) talk 11:58, 5 October 2011 (UTC)
  • I want it back - as Bushranger says, it is necessary. I would also like an option "only show edits that are not latest revisions" - I use my own contribution history as a sort of watch list and this option would help greatly. But I will file a separate Bugzilla request for it. — RHaworth (talk · contribs) 02:05, 6 October 2011 (UTC)
The "show only edits that are not latest" option is available now with my HideTopContrib script. Mark Hurd (talk) 11:13, 6 October 2011 (UTC)
  • Definitely agree that we need this feature. I even find it useful when looking at my own contributions, e.g. when I'm trying to find the most recent time that I participated in AFD. Nyttend (talk) 03:36, 6 October 2011 (UTC)
I definitely want it back. Unless, as Will Beback said above, a good reason is given for its removal, it should be restored. Those of us who contribute in multiple namespaces find it not only useful, but necessary. ---RepublicanJacobiteTheFortyFive 03:41, 6 October 2011 (UTC)
Adding to the pile-on. This is a VERY useful feature for me, especially dealing with certain types of vandalism. I would like to see it back. Trusilver 03:44, 6 October 2011 (UTC)
Totally agreeing with everyone. It's useful for checking contributions, especially vandals'. One reason my own MediaWiki installations will stay on 1.17 for now.Jasper Deng (talk) 03:47, 6 October 2011 (UTC)
According to this comment on the bug, the feature was not removed from MediaWiki, but merely disabled for very large wikis like this one. jcgoble3 (talk) 04:21, 6 October 2011 (UTC)
I also agree, good for finding copyvios as well Secret account 03:53, 6 October 2011 (UTC)

Lets have it back, please. JORGENEV 04:09, 6 October 2011 (UTC)

I twenty-third the call for restoration. I used that feature nearly every day.--Fuhghettaboutit (talk) 04:15, 6 October 2011 (UTC)
Twenty-fourthed - ditto. JohnCD (talk) 09:39, 6 October 2011 (UTC)
I can't believe that someone has removed this feature. It's something I use nearly every day, especially if I want to remember what talk pages I've been posting on (even if they are not on my watchlist, which is too big anyway), this has got to be restored to working order as soon as possible. --Hibernian (talk) 04:39, 6 October 2011 (UTC)
Agreed with all, we need the namespace option back. Whose idea was it to remove it? --Bryce Wilson | talk 05:06, 6 October 2011 (UTC)
I'd like this functionality back as well. It is also useful to be able to see related changes and watchlists and other reports filtered by namespace. An incredibly useful feature. Why it was removed with no explanation or warning is worrying. It makes me feel that the Wikimedia software used on Wikipedia is that little bit more unstable and less reliable, and makes me wonder what else could just be removed like this. Carcharoth (talk) 05:12, 6 October 2011 (UTC)
Just one more voice with the same request--this is necessary. Just to give another rationale why, a complaint has just been raised on ANI about a problematic user, and I need to know if the user has ever discussed the problems on user or article talk pages (i.e., are they completely uncommunicative, or just generally so), in order to advise how to move forward (because completely uncommunicative in the face of complaints is highly block worthy, while the latter may indicate other approaches). Searching through all of the communications by hand is much more time consuming and harder to get the "big" picture. Qwyrxian (talk) 05:15, 6 October 2011 (UTC)
Namespace back, needed.--Musamies (talk) 05:25, 6 October 2011 (UTC)
In favor of restoring the namespace filtering feature if at all possible. Sorry for joining late. --Lexein (talk) 10:38, 6 October 2011 (UTC)
  • WP:SNOW close this already and turn the feature back on. It's obvious that there is consensus to restore the feature, without a single !vote in opposition to turning the feature back on. Now let's close this discussion already and re-enable the feature. This is seriously getting in the way of accomplishing real work on this encyclopedia of ours. SchuminWeb (Talk) 06:10, 6 October 2011 (UTC)

Domas requested that this be removed because the implementation of namespace searching was such that it could create queries (either accidentally or maliciously) that would tie up database server resources for minutes at a time, and that represents an unacceptably large burden. Now it is also true that most queries created by namespace filtering do not take very long. Obviously many people want functionality like this, and the developers have been listening. More likely than not there will eventually be a restored filter with some new limitations to address the cases that can create a large server burden. For example, a filter might be limited to only consider a user's 5000 most recent contributions, or something like that. The exact form of what limits might be imposed are still being discussed and depend heavily on technical considerations about what can be implemented efficiently. Dragons flight (talk) 06:17, 6 October 2011 (UTC)

  • So Domas fixed bug by disabling feature? Congratulations. I suggest to disable editing on all wikimedia projects - it will allow to close many bugs on bugzilla. Bulwersator (talk) 08:01, 6 October 2011 (UTC)
    • Surprisingly not all "bugs"/issues can't be magically fixed like you seem to think they can, making "trollish" remarks like "lets just disable editing although" don't help anything. Peachey88 (T · C) 08:27, 6 October 2011 (UTC)
Who would want to be looking up 5,000 edits every day? I think most people only wish to look for the last hundred at the most. --Hibernian (talk) 08:53, 6 October 2011 (UTC)
As I understand it, the problem with the current implementation is it searches back through, say 5000 edits, to find 20 in the namespace you've chosen. As I understand it there isn't an index on namespace, or at least not a usable one. (BTW I want the feature back too.) Mark Hurd (talk) 10:32, 6 October 2011 (UTC)
Though I would like the feature back, I'd prefer the servers not crash or slow to a crawl every time I do something unexpected. I'm sure the developers can figure out something reasonable. WLU (t) (c) Wikipedia's rules:simple/complex 11:01, 6 October 2011 (UTC)
It's completely inacceptable that this feature was modified. As an German wikinews admin I can say that I did not use the feature because of making fun (like other things e.g. delivering cat images on user talk pages) but because of I needed to do so, especially for fighting vandalism, what especially in small projects with only a small number of admins which cannot keep up with the recent changes all day but will to browse thorugh a large number of recent edits once a day or maybe only twice a week. No good idea, what was done. --Matthiasb (talk) 12:33, 6 October 2011 (UTC)
  • If the only reason it was disabled was because "it can be slow", it should be immediately turned back on - we do not kill the patient to cure the disease. I think it's safe to assume that users who use the feature by-and-large realize that it is expensive, and understand the delay. If it's affecting site-wide performance, then that is a little more understandable but would like it if we could explore ways to provide this feature to users by way of a usergroup or by limiting the number of revisions one can ask for at one time (whenever I use the feature, I set this fairly low so the query does not take too long). –xenotalk 13:22, 6 October 2011 (UTC)

Ridiculous. If the idea is to make it impossible to locate diffs, accomplished. This won't work. SandyGeorgia (Talk) 13:31, 6 October 2011 (UTC)


Buttons in editing bar don't work

  Resolved
 – fixed
Extended content

Since yesterday, every page loads with an error message   at the bottom of my browser (IE7). The error seems to manifest itself in the editing bar, e.g.     etc. None of the buttons in the bar work when you click on them. Voceditenore (talk) 13:13, 5 October 2011 (UTC)

  Works for me (Firefox 9.0a/Win7x64). Did you try force-reloading? Regards SoWhy 13:41, 5 October 2011 (UTC)
Yep, I tried it. Makes no difference. :( Voceditenore (talk) 13:55, 5 October 2011 (UTC)
I just started my VM with WinXP/IE7 and I confirm this in that browser. It also doesn't work in IE8 on Win7x64. JS compatibility library is on. Works fine in Firefox and Chrome. Regards SoWhy 14:10, 5 October 2011 (UTC)
I'm totally non-techno-savvy. :) What does "JS compatibility library is on" mean? Does it mean someone is working in this? I hope so because it's real pain. Voceditenore (talk) 16:50, 5 October 2011 (UTC)
There is an option in Special:Preferences under "Gadgets" called "The JavaScript Standard Library" which is designed for browsers that do not support Javascript 1.6 (like Internet Explorer 7). If you insist on using such an outdated browser, you should turn this on to ensure compatibility in your browser. That said, the library does not solve this problem though. Regards SoWhy 17:05, 5 October 2011 (UTC)
Fixed in r99097. It was a stray comma, no less ^^ . --Catrope (talk) 13:36, 6 October 2011 (UTC)

I'm still having this problem. I have cleared my cache and restarted the program since the announcement that it was resolved was made, but to no avail. LadyofShalott 23:41, 8 October 2011 (UTC)

Are you using IE7? If so, have you enabled “The Javascript Standard Library” in Gadgets as Roan suggests? — MarkAHershberger(talk) 00:02, 12 October 2011 (UTC)

"m" and "b" edit indicators aren't bold

Makes them much harder to notice. –Drilnoth (T/C) 13:19, 5 October 2011 (UTC)

  Resolved
 – fixed
Extended content
You can use this in your skin.css (e.g. Special:Mypage/vector.css) or Special:Mypage/common.css (to have it in all skins) to regain that feature for you:
/* Re-bold-en minor and bot edits in contributions, history, recent changes */
abbr.minoredit, abbr.botedit {
  font-weight: bold;
}
Regards SoWhy 13:53, 5 October 2011 (UTC)
Was this a purposeful change? Seems counter-intuitive to me. ♫ Melodia Chaconne ♫ (talk) 14:02, 5 October 2011 (UTC)
Ditto. —{|Retro00064|☎talk|✍contribs|} 22:26, 5 October 2011 (UTC)
Thanks for this. I also added it to my vector.css and it works fine. Toshio Yamaguchi (talk) 14:03, 5 October 2011 (UTC)
Many thanks. Put it in common.css and my MonoBook is now bolded! :) - The Bushranger One ping only 19:31, 5 October 2011 (UTC)
Yup, adding that code to monobook.css fixed it for me also. Thanks! :-) —{|Retro00064|☎talk|✍contribs|} 22:24, 5 October 2011 (UTC)
This seems to be coming and going on a regular basis. I can no longer find the relevant CSS in trunk. How about putting it in common.css? Edokter (talk) — 14:42, 5 October 2011 (UTC)
Fixed now, see r99089. --Catrope (talk) 13:34, 6 October 2011 (UTC)

Monobook skin loses format on 3rd display

  Resolved
Extended content

I am using the monobook skin (which is remembered from Special:Preferences settings). However, after displaying 2 pages, the 3rd page (such as edit-preview) loses the format and shows the page in list format (with menu items listed down the page under the edit-box). At that point, the screen format is corrupted and the Main-page hyperlink area extends into the edit-box and clicks to the main-page when trying to click on text in the edit-window. This is, by far, the worst corruption of the English Wikipedia since the dawn of time. It is a good time to talk about stopping future updates to Wikipedia: it can no longer function when updating to MediaWiki 1.18 or such. It is time to question whether trivial improvements in other areas are worth utterly destroying the edit-window for experienced users. I think it is not: it is high time to place an announcement:

Upgrading to MW 1.18x, Wikipedia will be basically unusable this week

I wish there was an alternate, stable Wikipedia which could edit the same files, to allow users to make progress despite all the useless updates to the MediaWiki software. Sometimes, pages can be copied to Simple Wikipedia for edit-preview, then copied back after the edit-changes have been tested. However, recently all other-language Wikipedias seem to get update-corrupted within days of each other. WP has become too complex to be updated without breaking for thousands of users. -Wikid77 (talk) 15:22, 5 October 2011 (UTC)

This was caused by an oversight where someone forgot to disable three lines of code that made ResourceLoader slow (a lesson we learned during the 1.17 deployment but was somehow missed for 1.18), should be fixed now. --Catrope (talk) 17:35, 5 October 2011 (UTC)
  Resolved
 – duplicate of #Classic skin sidebar
Extended content

Using Monobook on a WinXP computer and IE7, the different sections in the left sidebar are significantly distant from each other (5.5 cm, roughly). This is artificially increasing the length of smaller pages and taking the "interaction" and "toolbox" sections below the fold. Risker (talk) 15:22, 5 October 2011 (UTC)

You probably need to clear your browser cache. —TheDJ (talkcontribs) 16:07, 5 October 2011 (UTC)
Did it before posting here, did it again before responding. No change. Risker (talk) 16:26, 5 October 2011 (UTC)
Have you tried disabling your gadgets and blanking Special:MyPage/monobook.js ? —TheDJ (talkcontribs) 16:43, 5 October 2011 (UTC)
Turned off all my gadgets, blanked my .js, logged out, cleared my caches, logged back in and .... no change. (I do appreciate the suggestions.) Risker (talk) 17:02, 5 October 2011 (UTC)
Have the same problem with IE 8. --Túrelio (talk) 06:37, 6 October 2011 (UTC)

Special:AllMessages

  Resolved
 – fixed
Extended content

The row highlighting is missing from Special:AllMessages. ---— Gadget850 (Ed) talk 15:56, 5 October 2011 (UTC)

Highlighting works for me, but the green/red backgrounds are now gone. Edokter (talk) — 16:08, 5 October 2011 (UTC)
bugzilla:28425TheDJ (talkcontribs) 16:10, 5 October 2011 (UTC)
You mean bugzilla:31380 :) Edokter (talk) — 16:35, 5 October 2011 (UTC)
Yes, I meant the background colors. More difficult to pick out changes. ---— Gadget850 (Ed) talk 16:44, 5 October 2011 (UTC)

Fix was deployed earlier today. --brion (talk) 01:44, 6 October 2011 (UTC)

All scripts broken in IE

  Resolved
 – fixed
Extended content

I'm seeing these types of error messages in the IE (9) console (press F12): "SEC7112: Script from http://en.wikipedia.org/w/index.php?title=User:Shubinator/monobook.js&action=raw was blocked due to mime type mismatch". Seems like something similar to this. Shubinator (talk) 17:02, 5 October 2011 (UTC)

As intended, won't change due to Bugzilla:15461 being fixed in 1.18. If you use action=raw, you should also use &ctype=text/javascript — ☠MarkAHershberger☣ (talk) 00:29, 6 October 2011 (UTC)
I've actually found the source of the missing ctype parameter: Special:MyPage doesn't pass it on when redirecting to the actual userpage. Shubinator, you should be able to work around this by changing User:Shubinator/vector.js to reference User:Shubinator/monobook.js directly instead of Special:MyPage/monobook.js. --brion (talk) 00:56, 6 October 2011 (UTC)
Worked like a charm, thanks! :) Shubinator (talk) 03:31, 6 October 2011 (UTC)
Proper fix for this is ready, should get deployed somewhere soonish. --brion (talk) 01:17, 6 October 2011 (UTC)
Fix has been deployed. --Catrope (talk) 13:30, 6 October 2011 (UTC)

Special:Contributions formatting

  Resolved
 – invalid: Please report any issues with this gadget here
Extended content

I'm using Firefox 7.0.1 with the Vector skin. I've already completely cleared the cache and restarted the browser. There is some kind of issue with the Special:Contributions page. Under the "User contributions" title at the top of the page, it says "For [UserFoo]' (talk | block | block log | uploads | logs | deleted user contributions | user rights management | filter log) • Javascript-enhanced contributions lookup 0.2 enabled. You may enter a CIDR range or append an asterisk to do a prefix search." That last sentence is unusual because (1) it overlaps the "Help:User contributions" link at the right side of the screen, and (2) I'm looking at a registered account, not an IP address, so why the info about CIDR range? (Maybe it was always like this and I just noticed) -- Gyrofrog (talk) 17:51, 5 October 2011 (UTC)

References in galleries

  Resolved
 – fixed
Extended content

In an old version of Elizabeth II,[22] there are <ref> markers in a <gallery> that are broken. DrKiernan (talk) 18:53, 5 October 2011 (UTC)

This is almost certainly a different manifestation of the same nested ref bug described above and in bugzilla:31374. It is likely to hit most forms of nesting. Dragons flight (talk) 21:18, 5 October 2011 (UTC)
It is under discussion on Bugzilla. ---— Gadget850 (Ed) talk 21:21, 5 October 2011 (UTC)
  Resolved
 – fixed
Extended content

Now that you can click on the column titles to sort (rather than just the icon) any links contained therein stop working. This is most problematic when it's a notes link, such as that at 2010 World Monuments Watch. violet/riga [talk] 19:18, 5 October 2011 (UTC)

Gah! Between this and the comments above about sortable tables, it really makes me wonder. Why was it changed at all? It was working just fine before and in a way that people would expect it to work. –Drilnoth (T/C) 19:23, 5 October 2011 (UTC)

This part of the issue should be fixed. Dragons flight (talk) 20:52, 5 October 2011 (UTC)

Excellent. –Drilnoth (T/C) 21:45, 5 October 2011 (UTC)

Overwritten file still shows old version

  Resolved

After uploading a series of project assessment maps, the old versions which were overwritten still show on the file's page. (example) However, thumbnails show the correct version, and if you click the image, the current version shows. –Fredddie 21:56, 5 October 2011 (UTC)

bugzilla:31382 " Fixed and live @ 2011-10-05 20:22:30 UTC. Confirmed by uploading a new file version over a test file. New file versions uploaded between 02:04 and 20:22:30 UTC (could) need a manual purge to update to the new file version correctly" —TheDJ (talkcontribs) 23:34, 5 October 2011 (UTC)

If you can make IE crash pretty reliably we need your help.

  Resolved
 – fix being deployed likely cause was jquery 1.6.2

See next section. Comment on the the bug or leave a comment here. (Comments on the bug will be seen sooner, though.) — MarkAHershberger 20:02, 6 October 2011 (UTC)

XP, IE8, Vector skin. Mostly it just refreshes the tab, displaying a popup thingy for a few seconds. Sometimes, IE decides this is happening too often and gives up, doing exactly what Chris describes in Comment 7 on the bug report. This morning, for the first time, IE said the site was misbehaving and that IE needed to close down. It proposed to send a bug report to MS, which I permitted, and then IE did NOT close (this is quite unusual). What (non-technical) help do you need? --Stfg (talk) 09:32, 7 October 2011 (UTC)
Some more details: (1) After opening a brand new browser window and loading any WP page, the first time I click My Watchlist, IE doesn't complain. (So far, My Watchlist is the only page with this property). On all subsequent occasions in that window+tab, it complains even for My Watchlist. (2) It happens not only on clicking Wikilinks, but also when loading WP pages from IE Favourites. (3)It doesn't happen the first time a WP page is loaded into a new IE tab. But after it has happened once in any tab, it seems to be there for life: you can load pages from other sites for as long as you like, then then if you try to load a WP page again after that, the problem is still there. Deleting temporary internet files and history does seem to reset the tab, though. --Stfg (talk) 10:20, 7 October 2011 (UTC)

() It's wrong to say that IE8 is crashing, and this is leading some people on Bugzilla to say it's MJ$ not MW. Although there have apparently been some crashes (and referencing the zero address is certainly a M$ bug), the showstopper is the one where a "website problem" necessitates IE refreshing the tab, which it quite often gives up on. The browser does NOT crash, and if it doesn't give up then it does load the requested page. But when it gives up, if you're editing, you can waste work and get in a mess. The problem doesn't arise, by the way, with Classic skin, so there's something in that one. But Classic skin is hopeless. For me, this is a complete show-stopper. I've suspended all but trivial editing, and won't be trying show-preview on this one either. --Stfg (talk) 21:43, 7 October 2011 (UTC)

The "This tab has been recovered" error in Internet Explorer 8 seems to only come up for me when I get on Wikipedia from another website. If I go to Wikipedia using the Run dialog (which starts Internet Explorer if it is not already running; otherwise, it adds a new tab to the existing one), then I do not get the error. I do not seem to get the error when browsing or editing Wikipedia. I am using I. E. 8 and Windows Vista 32-bit, and use the Monobook skin here on Wikipedia. Regards, —{|Retro00064|☎talk|✍contribs|} 22:00, 7 October 2011 (UTC).

Special pages

  Resolved
 – fixed
Extended content

Through 1.17, the message at the top of special pages, such as Special:DoubleRedirects, said when the page contents had last been updated. Now, under 1.18, it just says "The following information is cached and may not be completely up to date." Why was this done? Providing the timestamp for the last update can't be much of a resource burden, and it's information that has some use to some users. --R'n'B (call me Russ) 11:23, 6 October 2011 (UTC)

This was a bug. The timestamp is back now. --Catrope (talk) 13:27, 6 October 2011 (UTC)


  Resolved
 – Fixed but not yet deployed —TheDJ (talkcontribs) 14:23, 8 October 2011 (UTC)

Seems broken, I use this to limit the pages I watchlist (normally last 500 + about 100 that I always want on the list) When I open it (FF 3.6) I see &quot; instead of " in page titles. ΔT The only constant 16:05, 6 October 2011 (UTC)

Reported as bugzilla:31435TheDJ (talkcontribs) 16:21, 6 October 2011 (UTC)


Crashing Internet Explorer

  Resolved
 – fix being deployed and duplicate of #Problem with Internet Explorer
Extended content
  • My Internet Explorer (Windows XP Media edition) has gone haywire since yesterday, or should I say, since 1.18. Normal editing and saving is impossible with the following message popping up with every next move. For example, in order to post this message, I had to click Save without Preview and recover it later, because the page crashed again twice:
File:IE-needs-to-close-1.jpg

When I click off the above dialog box "Send" or "Don't Send" the screen is restored by IE to its initial stage with the following message: "This tab has been recovered." And naturally, all new stuff in the edit box disapears without the trace when the page is recovered. -- A. Kupicki (talk) 23:05, 6 October 2011 (UTC)

Though I usually use firefox, I used IE 8 yesterday to edit due to problems from 1.18 and faced the same problem. --Redtigerxyz Talk 05:32, 7 October 2011 (UTC)
All yesterday and again today several situations have caused a Microsoft "IE needs to close" message to appear. IE 8 is what I am using & Windows XP. If save has been pressed before that messge appears saving still happens very slowly if "don't send" is selected. However after a number of these incidents the attempt to reopen the tab fails and MediaWiki stops operating completely. It will work again if opened again in another tab but usually not for very long before another breakdown.--Felix Folio Secundus (talk) 10:13, 7 October 2011 (UTC)
Please note, the new screenshot shows exactly what happens when the 'Save page' button is pressed after editing. No improvement since yesterday. -- A. Kupicki (talk) 15:11, 7 October 2011 (UTC)

Please see #Problem with Internet Explorer. Bug is reported to the server admins/devs. --Saibo (Δ) 17:21, 7 October 2011 (UTC)

Sorry for the inconvenience! We've been able to reproduce this issue in some IE8 configurations and are currently investigating a number of possible fixes.--Eloquence* 18:32, 7 October 2011 (UTC)

Parts of the toolbar seem shy

I have had about half the toolbar vanish on a number of occasions today. No big deal, but a singular occurrence.Rich Farmbrough, 15:53, 7 October 2011 (UTC).

A plural occurrence, surely ;) --Tagishsimon (talk) 16:50, 7 October 2011 (UTC)
Maybe the same as this problem on ptwiki? — MarkAHershberger 21:01, 8 October 2011 (UTC)
It sounds like the one I'm having - if I could understand what the heck that ptwiki thing was about, it might be that too. (Grumble time: Some of us haven't done any programming or coding since black & white COBOL on MS-DOS 3.3 - ah, the days of .BAT files...) Whatever this is is still going on here - the Twinkle side of the tools don't appear, and the compression thingy doesn't, and apart from at the CSD page, all the 'show/hide' buttons are missing, and the drawers all open, and the reason for deletion doesn't appear in the delete screen. Reloading the page usually gets things working again, but sometimes has to be Ctrl-Reload (or Shift-Reload - haven't needed that for a couple of days). It's damned annoying. Peridon (talk) 18:41, 10 October 2011 (UTC)
At least MS-DOS 3.3 actually worked as per specification, which is more than can be said of [redacted]. BTW my only formal qualification is in COBOL-85... but our terminals were black-and-amber <sigh>. Anyway. The toolbar. Since MW1.18 rollout, about half the time I don't get the whole thing, but just one button, the   "Insert Citation" one. I have Enable enhanced editing toolbar turned off, because the new version is even less reliable. It's definitely not exclusive to en.wp, because it also happens on commons (where I get none at all: commons doesn't seem to have Wikipedia:RefToolbar 1.0. Pressing the browser's refresh button (F5 in Firefox) sometimes loads the missing buttons. --Redrose64 (talk) 19:02, 10 October 2011 (UTC)
Oh yes the upgrade from black and green to black and amber! (And the one before from teletype to black and green.)
The "cite" button is noticeably higher than the rest. I presume a different mechanism puts that there. ~~
This is a breakage of the local modification script User:MarkS/extraeditbuttons.js. Unfortunately, it seems to be unmaintained. —TheDJ (talkcontribs) 20:14, 10 October 2011 (UTC)
As mentioned on bug 31511, the script was rewritten on de.wp and it works fine as long as we do not use it as a gadget marked as ResourceLoader-compatible (because it is not compatible, and the bug indicates that something needs to be changed in MediaWiki in order to that compatibility to be possible).
For now, we could just copy the script from de:MediaWiki:Gadget-Extra-Editbuttons.js and translate its button descriptions back to English, because it is not being loaded as a gadget and thus the bug doesn't affect it. Helder 20:31, 10 October 2011 (UTC)
(Noticeably it seems to be Chrome where I have the problem.) That sounds cool. Where should the result go? Rich Farmbrough, 22:57, 10 October 2011 (UTC).

Show MediaWiki messages

"To find the message responsible for producing a particular piece of interface text in MediaWiki 1.18 and above, change to the special 'language' code qqx by appending ?uselang=qqx to the URL. All the messages will then be replaced by their message keys, so you can identify which message is responsible."

This is a really neat feature for advanced editors. I created a bookmarklet that will open the current page in a new tab, appending ?uselang=qqx. MediaWiki will not create a proper drag & drop link, so copy this:

javascript:window.open('?&uselang=qqx');void%200

Paste it as a bookmark and name it "Wikipedia qqx" or whatever you desire. ---— Gadget850 (Ed) talk 13:44, 9 October 2011 (UTC)

Really neat indeed. If the url already contains ? then &uselang=qqx without a ? works in the url, but it doesn't work for me in a Firefox bookmarklet if ? is simply removed from your javascript. How can this be made to work? PrimeHunter (talk) 14:21, 9 October 2011 (UTC)
Example please. I had test this with What the Bleep Do We Know!?. ---— Gadget850 (Ed) talk 14:31, 9 October 2011 (UTC)
Example: http://en.wikipedia.org/w/index.php?title=Kitten&action=history. This works: http://en.wikipedia.org/w/index.php?title=Kitten&action=history&uselang=qqx. A second question mark breaks it. PrimeHunter (talk) 14:37, 9 October 2011 (UTC)
Yes; and I see how that would be useful. I don't know how to fix that as I am still learning this stuff. ---— Gadget850 (Ed) talk 14:45, 9 October 2011 (UTC)
Until a JS guru comes up with something, best I can figure is a separate bookmarklet without the ?. ---— Gadget850 (Ed) talk 18:14, 9 October 2011 (UTC)
Now available as a Gadget™ under the Advanced section. Edokter (talk) — 18:45, 9 October 2011 (UTC)
Just saw the update on my watch list. Sweet. Thanks. ---— Gadget™850 (Ed) talk 20:22, 9 October 2011 (UTC)
You can try
javascript:window.open(location.href+(location.search?"&":"?")+"uselang=qqx","_blank");
Helder 19:52, 9 October 2011 (UTC)
Thanks. That works on all pages I have tried. The Gadget under the Advanced section doesn't work on url's with a question mark such as page histories and edit pages. PrimeHunter (talk) 16:49, 10 October 2011 (UTC)
Trying to fix the gadget, but I'm puzzled; I see an empty search method... what does it search for? Edokter (talk) — 18:54, 10 October 2011 (UTC)
Never mind, I figured it out. Edokter (talk) — 18:59, 10 October 2011 (UTC)
For the record: see location.search. Helder 19:46, 10 October 2011 (UTC)
And it is fixed before I could even bring it up. And there was much rejoicing. ---— Gadget850 (Ed) talk 21:12, 10 October 2011 (UTC)

There is a problem with viewing the Robbie Kerr article in that it overlaps the left-hand navigation links in the Monobook skin rather than creating a scroll bar and hanging off the right edge of the window. Using Firefox 3.6.23 on XP. It is OK if the window is widened. Keith D (talk) 22:58, 9 October 2011 (UTC)

Fixed. Someone had copied in some elements that didn't really belong into that article and messed up the formatting. —TheDJ (talkcontribs) 23:18, 9 October 2011 (UTC)
Thanks. I had not spotted that. Keith D (talk) 13:04, 10 October 2011 (UTC)

Server error (again)

I keep getting the following error message when I attempt a contributions search in the Wikipedia Talk namespace:

Proxy Error
The proxy server received an invalid response from an upstream server.
The proxy server could not handle the request GET /wikipedia/en/w/index.php.
Reason: Error reading from remote server
Apache/2.2.8 (Ubuntu) mod_fastcgi/2.4.6 PHP/5.2.4-2ubuntu5.12wm1 with Suhosin-Patch mod_ssl/2.2.8 OpenSSL/0.9.8g Server at secure.wikimedia.org Port 443


I reported this problem earlier this month, was advised to use the new HTTPS server, and the problem stopped. But, in the last 3 days (at least) the problem has returned. Anyone have any ideas about this? Is this related to the other bugs in the new system? ---RepublicanJacobiteTheFortyFive 17:57, 10 October 2011 (UTC)

"GET /wikipedia/en/w/index.php." and "secure.wikimedia.org" so you are using the old server again. :D This is not that strange, we have MANY existing links to the old https server, and it's all a bit confusing at the moment. Soon the old server will be taken out of commission, and all links pointing to it will go trough a redirecting script that will make it point to the new https servers. —TheDJ (talkcontribs) 20:00, 10 October 2011 (UTC)
Thank you for the response, DJ, but I am still perplexed. I guess this will all shake out in the end. Cheers! ---RepublicanJacobiteTheFortyFive 01:22, 11 October 2011 (UTC)

Delete delays

Clicking the delete button in the delete screen itself is currently often followed by a very long delay - long enough to go back to CSD, select another candidate, load it up, make a decision and try to delete it. When this happens, everything else seems to be at normal speed. Today, I got a Database notice instead of an Action completed - but didn't write it down. Will if it does it again. Anyone else getting this? Peridon (talk) 19:29, 10 October 2011 (UTC)

Everything seems slower since the last update, and yes, delete is often very slow now.  Ronhjones  (Talk) 23:50, 10 October 2011 (UTC)

Problems in be.wiki

Because, unfortunately, in our language section, no person who would be able to fix those troubles that we delivered a new version MediaWiki, I will write here about the current problems with great hope that someone who can help will see it and to do something.

From the ground: we no longer updated statistics of FlaggedRevs and hotkat works only to add categories.

Those who want to help: please write on my page talk to be.wiki. You can both Russian or English. Many thanks in advance--Хомелка (talk) 19:56, 10 October 2011 (UTC)

I've made a suggestion at be:MediaWiki talk:Gadget-HotCat.js#Load_from_Commons. I hope that helps. Helder 20:46, 10 October 2011 (UTC)

Page too big to load

  Resolved

(Also posted at WP:AN) According to the NFCC#9 violation report, the file User:Multichill/Free_uploads has numerous non-free images in userspace which need to be removed. The problem is, the page appears to be so big that I can't load it. Can anyone else? And if so, could someone remove the following files from it?

Thanks, Black Kite (t) (c) 23:54, 10 October 2011 (UTC)

Seems to be a memory issue. Couldn't the page be entirely deleted?--Bbb23 (talk) 00:03, 11 October 2011 (UTC)
It is updated by User:BotMultichillT; that bot might need to be modified to remove the oldest subpage from the list when it adds another. The page in question transcludes tons of other pages by date; you should look there. Also, it should be possible to view the history of, edit, or delete that page by manually navigating to the history page (e.g. to remove the old subpages that make the page too big to load). PleaseStand (talk) 00:12, 11 October 2011 (UTC)
I think the Multichill bot has not been working. DASHbot has removed some, but not all - i.e. [one missed but one removed here. However, I've managed to remove all the offending images from the subpages, so all is good now. Black Kite (t) (c) 00:33, 11 October 2011 (UTC)

Wikipedia articles not on Wayback

I find Wikipedia's co-founder Larry Sanger complaining on Twitter that this page prevents old versions of Wikipedia articles from appearing on Wayback. (BTW, it occurs to me that maybe persons of youngness might fail to recognize the obvious humorous allusion to Rocky and Bullwinkle....) The concern expressed in the file is Wikipedia user pages, but apparently it affects everything on Wikipedia including articles. Should this be done something about? Michael Hardy (talk) 01:14, 7 October 2011 (UTC)

No, we regularly upload full database dumps to archive.org. which is far far better than crawling the site. ΔT The only constant 01:16, 7 October 2011 (UTC)
There is another person also complaining at the archive.org forum about this problem for another website. Perhaps archive.org changed their configuration ? —TheDJ (talkcontribs) 12:25, 7 October 2011 (UTC)
I am looking for a deleted article, even though I regularly d/l dumps, it would be far easier to use Wayback than trawl through old dumps, possibly having to spend 8 hours grabbing each, and about an hour search time for every refinement of the search terms. But finding Sanger complaining about Wikipedia on Twitter is hardly a surprise.   Rich Farmbrough, 16:15, 7 October 2011 (UTC).
Indeed, I make many quite legitimate complaints about Wikipedia on Twitter. You're welcome to follow me if you find the thought amusing.
OK, I'll simply ask: why was there only one snapshot of http://en.wikipedia.org/ in July, and why was that the last one? http://wayback.archive.org/web/*/http://en.wikipedia.org --Larry Sanger (talk) 19:38, 7 October 2011 (UTC)
Huh--the problem may be with archive.org. http://wayback.archive.org/web/*/http://www.citizendium.org --Larry Sanger (talk) 19:48, 7 October 2011 (UTC)
NOW, I find the thought amusing :D sorry, couldn't help saying that. —TheDJ (talkcontribs) 20:28, 7 October 2011 (UTC)
What I find amusing is that I never see your name without the phrase "complaining abut Wikipedia" somewhere near, not the actual complaints themselves. Rich Farmbrough, 22:50, 10 October 2011 (UTC).
What bit of the robots.txt looks like it excludes the Internet Archive bot? Only reference I see for ia_archiver is commented out, and that's to try to exclude the User namespaces only. --brion (talk) 00:55, 8 October 2011 (UTC)
I don't know the answer, Brion, but look at http://web.archive.org/web/20110721203834/http://en.wikipedia.org/w/index.php . I'm just quoting archive.org: "Page cannot be crawled or displayed due to robots.txt." That's what made me think there was something wrong with robots.txt. But I don't know what bit excludes the bot, if any. The problem indeed could be with archive.org, but right now, it looks like nobody knows why that "page cannot be crawled" message is appearing. --Larry Sanger (talk) 01:04, 8 October 2011 (UTC)
We exclude everything of the form /w/ from crawling because those are for active scripts. You should look for pages of the form http://en.wikipedia.org/wiki/ ... As I recall (from a couple years ago), the Internet Archive used to have a multi-month delay between when something would be crawled and when it would appear in their online archive. If that is still the case then it could explain the lack of recent records. Dragons flight (talk) 04:58, 8 October 2011 (UTC)
Well, that completely clears that up. I'll post a Twitter correction. --Larry Sanger (talk) 18:57, 8 October 2011 (UTC)
See http://www.archive.org/about/faqs.php#103: "Why are there no recent archives in the Wayback Machine?" PrimeHunter (talk) 11:32, 11 October 2011 (UTC)

User:Allstarecho/autoarchive.js not working for me.

I recently found the above script on Wikipedia and thought it would be very useful for my purposes. I added the relevant command to my common.js page. Now I have the "Auto-Archive" button in the top right-hand corner next to my user name. When I click it while on an article talk page though, I am redirected to the page where one makes edits and nothing is added there by the script. Would someone be willing to figure out what problems, if any, there are in this script. Unfortunately, I can't understand Javascript. Help will be greatly appreciated. I am using Firefox v7.0 -XJDHDR (talk) 16:27, 10 October 2011 (UTC)

The auto archive actually only place a configuration template on the edit page as shown below. The you need to have a bot, pywikipedia bot for example, to do the archiving for you. --Kurniasan (talk) 17:20, 10 October 2011 (UTC)
{{talkheader}}
{{User:MiszaBot/config
|archiveheader = {{Talkarchivenav}}
|maxarchivesize = 80K
|counter = 1
|minthreadsleft = 1
|minthreadstoarchive = 1
|algo = old(15d)
|archive = Talk:PAGE NAME/Archive %(counter)d
}}
Thank you for the reply. The problem I see though is that the script does not add the above template to the talk page. As a matter of fact, it does not add anything to the talk page, not even a single character. -XJDHDR (talk) 19:48, 10 October 2011 (UTC)
I'm sorry for misunderstanding. My common.js is a copy of yours, but the scripts seem to work fine in both Firefox 7.0 and Google Chrome. Also, I don't see any errors in the console. --Kurniasan (talk) 17:53, 11 October 2011 (UTC)

Can't view Wiki in Firefox

  Resolved
 – Disabled plugins
Extended content

I have used Firefox on Wiki for years with no problems BUT now when I go to a Wiki page everything appears normal for 1-2 seconds the suddenly all the text on the page collapses, leaving only the article title and the Rate This Page box. This only occurs on the English Wiki. Other languages load the page fine. I do not have this problem with the English Wiki on IE so its Firefox for sure. I have cleared the Firefox cache and also tried uninstall of Firefox and then a reboot of computer and new download of Firefox, but the problem persists. Help!--KeithbobTalk 21:58, 10 October 2011 (UTC)

Which version of Firefox, also which operating system are you using? I'm using Firefox 3.6.23 under Windows XP without the trouble you describe. --Redrose64 (talk) 22:31, 10 October 2011 (UTC)
Which skin are you using? Does it happen when you are logged out? PrimeHunter (talk) 22:42, 10 October 2011 (UTC)
I am using FF version 3.6 and Windows XP. The problem occurs whether I am logged in or logged out. So the skin is not the deciding factor. thanks for your help. --KeithbobTalk 00:12, 11 October 2011 (UTC)
I don't know what causes your problem but here are some things you might try: Make sure you clear the entire cache in Firefox. Get a newer Firefox version. I have the current stable release 7.0.1. Go to Special:Preferences#mw-prefsection-rendering and enable "Don't show the Article feedback widget on pages". Say whether http://en.wikipedia.org/wiki/Wikipedia:Village_pump_%28technical%29?useskin=monobook or http://en.wikipedia.org/wiki/Wikipedia:Village_pump_%28technical%29?uselang=qqx works (the latter will look weird outside the editable area, I'm not suggesting you use it permanently). Disable JavaScript in Firefox (at least long enough to tell us whether it works, see [23]). PrimeHunter (talk) 01:36, 11 October 2011 (UTC)
Actually, I was mistaken, I have Firefox 7.0.1 version. Clearing the cache did not help but I disabled all my Firefox add-on's and that seems to have cleared up the problem for now. Thank you very much for your assistance. --KeithbobTalk 23:59, 11 October 2011 (UTC)

Wikitable Sortable not working

  Resolved

Hi,

here is a table where I can't see any sort arrows, it worked fine before the software update. Anyone knows why?

Note that I read the previous thread above about the same issue and I didn't see anything that fixes my problem. For example, I can see the sort arrows in this page: List of Manchester United F.C. players. Tachfin (talk) 01:09, 11 October 2011 (UTC)

Look up a ways - section 2.10. Ravensfire (talk) 01:16, 11 October 2011 (UTC)
Can you be more specific? I've replaced Background with Background-color and purged my cache, and it didn't fix the problem . --Tachfin (talk) 01:24, 11 October 2011 (UTC)
Your table isn't specifying column headers. Instead you are FAKING headers with custom styling. The table sorter will no longer support this. Nor should it ever have supported it, it's bad semantics. See also Table help, specifically the part that talks about column headers. —TheDJ (talkcontribs) 10:19, 11 October 2011 (UTC)
Thank you for the fix. This table format was made by "the excel to wiki" converter Tachfin (talk) 10:33, 11 October 2011 (UTC)

Random portal components

{{Random portal component}} is occasionally returning subpage 0, even though it should always be >= 1 (see Template talk:Random portal component#Counting from 1 or from 0?). Since the talk pages of the relevant templates are rarely visited, I thought I'd bring it up here. It seems likely that the problem is with {{mod}} (returning -1), but I can't see what the problem is. Any assistance would be greatly appreciated. --Stemonitis (talk) 08:29, 11 October 2011 (UTC)

  Resolved
--Stemonitis (talk) 13:14, 11 October 2011 (UTC)

Weird watchlist

Extended content

I just checked the list of pages on my watchlist and I spotted Hurricane Shit (2011) and Hurricane Shithead, even though I never watchlisted or visited them. Why did they appear? HurricaneFan25 13:35, 11 October 2011 (UTC)

You were watching Hurricane Irene (2011), when vandals moved it to the new titles. See Help:Watching pages#Moves, creations and deletions. The vandalism has since been undone and concealed, presumably to deny recognition, but your watchlist will still show the effects. --Stemonitis (talk) 13:42, 11 October 2011 (UTC)
(edit conflict) The two pages you mention existed in the past but have been deleted (see Hurricane_Shit_(2011) and Hurricane_Shithead). These pages are the result of Page-move vandalism where someone moved that page to the title you see on your watchlist. I guess you mean your raw watchlist. Deleted pages remain visible in your raw watchlist until you remove them from there. Toshio Yamaguchi (talk) 13:52, 11 October 2011 (UTC)
Added this as a comment on Page move vandalism creates watchlist clutter. Hope to bump its priority for a fix. — MarkAHershberger(talk) 19:59, 11 October 2011 (UTC)

Error when attempting to create new book

Extended content

When I attempt to create a new book I get this error:

Unexpected non-MediaWiki exception encountered, of type "UsageException"
badtoken: Invalid token

#0 /usr/local/apache/common-local/php-1.18/includes/api/ApiBase.php(1205): ApiBase->dieUsage('Invalid token', 'badtoken')
#1 /usr/local/apache/common-local/php-1.18/includes/api/ApiMain.php(590): ApiBase->dieUsageMsg('sessionfailure')
#2 /usr/local/apache/common-local/php-1.18/includes/api/ApiMain.php(678): ApiMain->setupModule()
#3 /usr/local/apache/common-local/php-1.18/includes/api/ApiMain.php(340): ApiMain->executeAction()
#4 /usr/local/apache/common-local/php-1.18/extensions/Collection/Collection.body.php(867): ApiMain->execute()
#5 /usr/local/apache/common-local/php-1.18/extensions/Collection/Collection.body.php(225): SpecialCollection->saveCollection(Object(Title), false)
#6 /usr/local/apache/common-local/php-1.18/includes/SpecialPageFactory.php(460): SpecialCollection->execute(NULL)
#7 /usr/local/apache/common-local/php-1.18/includes/Wiki.php(224): SpecialPageFactory::executePath(Object(Title), Object(RequestContext))
#8 /usr/local/apache/common-local/php-1.18/includes/Wiki.php(624): MediaWiki->performRequest()
#9 /usr/local/apache/common-local/php-1.18/includes/Wiki.php(531): MediaWiki->main()
#10 /usr/local/apache/common-local/php-1.18/index.php(57): MediaWiki->run()
#11 /usr/local/apache/common-local/live-1.5/index.php(3): require('/usr/local/apac...')
#12 {main}

At first I thought it might be an issue with Chrome, but I get the same error with Firefox too. If this isn't the right place for this kind of issue I apoligize. I didn't know where else to go. There isn't really an obvious way to report this kind of stuff on Wikipedia.--Sabre ball (talk) 15:28, 11 October 2011 (UTC)

Errors like this should be reported in Bugzilla. I've reported your issue here: stacktrace when creating a book. You can add your email as a CC to the bug to be updated on it when it is solved. Or you can just check in. -- MarkAHershberger(talk) 19:09, 11 October 2011 (UTC)

Wikipedia article traffic statistics

I created the article Seitajärvi 14 October 2010. Wikipedia article traffic statistics shows hits for before that, ie September, August etc 2010. User:Henrik hasn't edited since March 2011 so no point asking them about it. SlightSmile 02:49, 12 October 2011 (UTC)

It was probably linked to by other articles, and people occasionally clicked on the red links. ctzmsc3|talk 02:52, 12 October 2011 (UTC)

JS and DYK Check

Addressed Here and and Here, the DYK Check, Modern skin, Firefox and Windows XP. Cache purged. Still returns a Zero on the prose, with a pop up that says XML Error. --Maile66 (talk) 14:42, 9 October 2011 (UTC)

Do you still see this? What browser are you using? — MarkAHershberger(talk) 00:40, 12 October 2011 (UTC)
Yes, I still see this. Zero on the prose count. The XML Error pop up happens after I run the DYK check and click anything to change pages. Firefox 3.6.23, Windows XP --Maile66 (talk) 18:16, 12 October 2011 (UTC)

Asteroids game on Wikipedia?

Please take a look at User:LikeLakers2/asteroids/asteroids.js and User:LikeLakers2/asteroids2.js. I became aware of these after seeing a post by LikeLakers2 on User talk:Jimbo Wales. My inclination is to take them to MfD as not being an appropriate use of userspace or server resources. I'm not a programmer however, and wanted someone to tell why I'm completely wrong first, if I am wrong. LadyofShalott 11:55, 10 October 2011 (UTC)

Looks like it's derived from the code at http://kickassapp.com/ and is, in fact, a functional game. While this is clearly awesome, the precedent set for "find my hidden page" games quite obviously applies to fully-blown Asteroids clones. Chris Cunningham (user:thumperward) - talk 13:04, 10 October 2011 (UTC)
Server resources? It is javascript that runs off the client computer, not the server. Besides, the "highscore" thing is not off Wikipedia. That, and with the amount of traffic Wikipedia gets already, I don't think server resources is much of a problem. (my main reason it is from that site is because I didn't want the advertising showing for people other than me who are playing. It is basically that script with some lines of code removed. User:LikeLakers2/asteroids2.js is just a file that adds a link to it to play it. LikeLakers2 (talk | Sign my guestbook!) 15:04, 10 October 2011 (UTC)
It doesn't really matter how much; the server resources exist to build an encyclopedia, and this game is not involved in that. --Golbez (talk) 15:26, 10 October 2011 (UTC)
(e/c) You seem to have missed one - User:LikeLakers2/asteroids/excanvas.js. I haven't seen a user script game on Wikipedia before (agree with thumperward's "clearly awesome") and don't believe a user script like this has been taken to MfD in the past. --Philosopher Let us reason together. 15:28, 10 October 2011 (UTC)
Thanks for pointing that out. I'll be nominating this at MfD. LL2, NOT:WEBHOST. LadyofShalott 15:34, 10 October 2011 (UTC)
Poof. Gwen Gale speedily deleted them faster than I could finish the MfD. LadyofShalott 16:12, 10 October 2011 (UTC)
I've let LL2 know. While JS arcade games can be cool, scripts with no meaningful worth towards building the encyclopedia don't belong here and dealing with such content is most likely only a waste of volunteer time. Gwen Gale (talk) 16:35, 10 October 2011 (UTC)
What does 'CSD F10' have to do with these pages? Ruslik_Zero 18:39, 10 October 2011 (UTC)
"Files uploaded that are neither image, sound, nor video files, are not used in any article, and have no foreseeable encyclopedic use." Works for me. --Golbez (talk) 18:55, 10 October 2011 (UTC)
There are another 16,870 asteroid games here. Thincat (talk) 19:10, 10 October 2011 (UTC)
.js pages are not media files. They are user pages subject to Ux CSD criteria. Ruslik_Zero 19:14, 10 October 2011 (UTC)
You may have misread it, F10 has to do with the speedy deletion of non-media files. Gwen Gale (talk) 19:52, 10 October 2011 (UTC)

From what I know, F10 applies to filespace files only. LikeLakers2 (talk | Sign my guestbook!) 19:57, 10 October 2011 (UTC)

See WP:Wikilawyer (3 & 4). Gwen Gale (talk) 20:03, 10 October 2011 (UTC)
Since .js games are a vanishingly rare occurrence, I don't really see a problem with an WP:IAR "mis"-interpretation of F10 (though I would agree that it isn't actually F10). If you think they should be kept, the best thing to do is take it to WP:Deletion review. --Philosopher Let us reason together. 20:49, 10 October 2011 (UTC)
Since they were deleted out of process, I restored them. Ruslik_Zero 18:45, 11 October 2011 (UTC)
There were not deleted out of process as such and I'm aware of no consensus to restore them, nor of your having sought to gauge any consensus. Moreover, LL2 has not taken this to DRv. See also WP:BURO. Gwen Gale (talk) 19:06, 11 October 2011 (UTC)
In process/out of process... I'll go ahead and nominate them for MfD. LadyofShalott 19:22, 11 October 2011 (UTC)
If they're going to be restored for the time being, surely they all should be. User:LikeLakers2/asteroids/excanvas.js and User:LikeLakers2/asteroids.js are still redlinks. Jenks24 (talk) 19:38, 11 October 2011 (UTC)
Thanks. I restored User:LikeLakers2/asteroids/excanvas.js. The remaining one was only a redirect. It is at times difficult to find all deleted pages if they are deleted out of process. Ruslik_Zero 11:13, 12 October 2011 (UTC)
Truth be told, Ruslik, you're wholly mistaken, both as to finding deleted pages (all three are named in this thread and are also in my log), along with your notions of what's "in" or "out" of process and even what "process" is on en.WP. Gwen Gale (talk) 13:53, 12 October 2011 (UTC)

Please see Wikipedia:Miscellany for deletion/User:LikeLakers2/asteroids/asteroids.js. LadyofShalott 19:37, 11 October 2011 (UTC)

Wrapping in edit window.

For quite a long time, when hyphenated words hit the right-hand edge of the edit window, they have wrapped before the hyphen, like this:

double
-barrelled

This is just about bearable, but recently (possibly in 1.18, but I'm not sure) this behaviour has extended to commas, full stops, and I've also just seen "1st" wrap like this:

1
st

These are more troublesome when editing, because they neccessitate undesirable measures like temporarily inserting a newline to check that the existing text doesn't have incorrect spaces. I think a good fix would need to have the following attributes:

  • when a string is to be wrapped, any punctuation mark(s) should be placed on the first of the two lines;
  • strings incorporating digits followed by letters, such as ordinals, should not be wrapped;

(both of these excepting cases of force majeure).

(I use the standard edit window in Vector skin on Win XP). --Stfg (talk) 21:14, 11 October 2011 (UTC)

It also wraps "1½" like this:
1
½
--Stfg (talk) 21:28, 11 October 2011 (UTC)
I guess this is controlled by your browser and not Wikipedia. What is the browser? Does it happen at other websites? PrimeHunter (talk) 22:11, 11 October 2011 (UTC)
Thanks for your help, Primehunter. My browser is IE8 on Win XP. I don't type on many websites but, in one forum, hyphen, comma and period all behaved correctly, while 1st and 1½ wrapped the same as here. I'm beginning to understand how this could be the browser, if the edit box is simply an editable text control to which WP sends a text string for the browser to render. Is that how WP does it? The only thing is, I haven't changed the browser in any way recently, but the behaviour of sending commas and periods to the next line without the preceding word is new, I think. --Stfg (talk) 10:49, 12 October 2011 (UTC)
Currently, we don't do any manipulation of the text input control. See: Visual Editor for what we plan for the future. — MarkAHershberger(talk) 21:12, 12 October 2011 (UTC)
Thanks. --Stfg (talk) 22:00, 12 October 2011 (UTC)

Unlock

  Resolved

Letter removed from blacklist by User:Carnildo. --Philosopher Let us reason together. 22:36, 12 October 2011 (UTC)

Somebody please remove Mường Luân and Mường Lói from the spam filter black list I have some new articles ready.♦ Dr. Blofeld 22:09, 11 October 2011 (UTC)

The character ư is listed under "OBSCURE ASCII CHARACTER LOOKALIKES" at MediaWiki:Titleblacklist. It should remain there. If you create the pages in userspace without that character then an admin can move them for you. Or you can create them at Muong Luan and Muong Loi and request a move, since these titles should probably redirect anyway - if they shouldn't be the actual titles in the English Wikipedia per WP:COMMONNAME or Wikipedia:Naming conventions (Vietnamese). PrimeHunter (talk) 22:28, 11 October 2011 (UTC)

I am requesting that Muong Luan is moved to Mường Luân and Muong Loi to Mường Lói.♦ Dr. Blofeld 22:40, 11 October 2011 (UTC)

ư though is a letter in the Vietnamese alphabet. There are thousands of potential articles with that letter. SOmebody please consider taking this letter off the blacklist. There is many others with it in Template:Dien Bien Province. Or perhaps code it so Mường as a word is OK? ♦ Dr. Blofeld 22:41, 11 October 2011 (UTC)

I've moved the articles and copied this section to MediaWiki talk:Titleblacklist#Vietnamese for discussion on whether the letter should be removed from the Blacklist. --Philosopher Let us reason together. 22:50, 11 October 2011 (UTC)

I appreciate that, thanks Philosopher. ♦ Dr. Blofeld 22:52, 11 October 2011 (UTC)

__NOEDITSECTION__ stopped working

The __NOEDITSECTION__ command stopped working on my userpage. Any idea why? Dr.K. λogosπraxis 03:58, 12 October 2011 (UTC)

It does not work on my user page or on Wikipedia:Main Page alternative (text only) either. Perhaps this is a bug in 1.18? PleaseStand (talk) 04:17, 12 October 2011 (UTC)
This is what I am suspecting. Thanks. Dr.K. λogosπraxis 04:24, 12 October 2011 (UTC)
If I purge Peter Karlsson then the section edit links usually (not always) disappear. If I null edit it then they usually (not always) reappear. PrimeHunter (talk) 11:25, 12 October 2011 (UTC)
I've noticed that my talk page archives - all of which contain a {{talkarchive}} - now have section edit links where previously they didn't. In the absence of a |category= parameter (which I haven't used), this template adds __NOEDITSECTION__. --Redrose64 (talk) 11:43, 12 October 2011 (UTC)
I purged the user page once and the edit buttons are still there. I used to see a bug link for media wiki where one could report this but I don't think it is displayed any longer. Does anyone know where this can be reported? Dr.K. λogosπraxis 13:33, 12 October 2011 (UTC)
https://bugzilla.wikimedia.org – Adrignola talk 14:37, 12 October 2011 (UTC)
Thank you very much. Dr.K. λogosπraxis 15:54, 12 October 2011 (UTC)
This seems to be due to the fix for T33445. What happened there was that, if the page happened to be rendered in a context that doesn't display section edit links (e.g. a printable view), it would be saved that way in the parser cache. So later when the page was reloaded from the parser cache it would continue to lack the section edit links. The fix for that bug changed things to always save the page with section edit links into the parser cache, and then to remove those links when actually outputting the page if they aren't supposed to be there for that particular pageview.
But it appears that when the page is loaded from the parser cache, it ignores the possibility that links should be hidden due to __NOEDITSECTION__ and so fails to remove them. See T33647. Anomie 16:32, 12 October 2011 (UTC)
Thank you Anomie for the explanation. Let's hope they fix it. Dr.K. λogosπraxis 17:09, 12 October 2011 (UTC)
If you see that our fix to a reported bug has caused another problem, please reopen the bug report or open a new bug and refer to the earlier bug report. __NOEDITSECTION__ is ignored when the page is read from the parser cache has been reported and (with it) a possible fix is provided. — MarkAHershberger(talk) 21:00, 12 October 2011 (UTC)
Don't you ever test this stuff? Malleus Fatuorum 00:35, 13 October 2011 (UTC)
It appears that they do. On us. Dr.K. λogosπraxis 00:39, 13 October 2011 (UTC)
Perhaps that's consistent with the idea that if you pay peanuts you attract monkeys. Malleus Fatuorum 00:47, 13 October 2011 (UTC)
I shudder at the thought. Especially if they all have wrenches. :) Dr.K. λogosπraxis 00:58, 13 October 2011 (UTC)

Database disconnections

Reported: Intermittent "cannot contact the database server" on https://en.wikipedia.org/. Please make further comments there. -- MarkAHershberger(talk) 18:16, 12 October 2011 (UTC)

Extended content

I have been recently experiencing problems on Wikipedia where an error page shows up instead of Wikipedia's real pages because the connection to MySQL failed. Usually a refresh does the trick, but sometimes many refreshes get the same thing.Jasper Deng (talk) 03:58, 12 October 2011 (UTC)

I've been getting that note too - and I've been getting it off and on all day - the one I'm getting has a "(Can't contact the database server: Can't connect to local MySQL server through socket '/var/run/mysqld/mysqld.sock' (2) (localhost))" on it. I'm using Firefox 7.01 on Windows 7 if that makes a difference. --Philosopher Let us reason together. 06:30, 12 October 2011 (UTC)

Special:Badtitle

If you enter an invalid title in the browser's url (for example, http://en.wikipedia.org/wiki/France} with the trailing brace), the software will take you to Special:Badtitle, which does not appear to exist. A "No such special page" error is then displayed instead of a "bad title" error. 28bytes (talk) 15:47, 12 October 2011 (UTC)

Yes, I believe the developers are aware of that. It's been mentioned before, but I'm not sure if a bug has been filed. It used the non-existent Special:Badtitle before too, but it seems to have broken due to other changes. MediaWiki could really use a less hacky way of displaying bad title errors. Reach Out to the Truth 16:04, 12 October 2011 (UTC)

Invert selection button missing

  Resolved

On the "View and restore deleted pages" the button for "Invert selection" appears to have disappeared since the last time I used the restore function. Rather frustrating when there are a lot of versions to restore. Keith D (talk) 18:04, 12 October 2011 (UTC)

MediaWiki now provides "Invert selection" button next to "Restore" button so local enwiki script was removed. FYI you can also use shift-clicking to select/deselect a range of checkboxes. — AlexSm 18:26, 12 October 2011 (UTC)
?? I don't see any invert selection button, just "Restore" and "Select all". –xenotalk 18:57, 12 October 2011 (UTC)
Just believe me and do not believe the text on that button :) Or you could delete MediaWiki:Undeleteinvert (now "Select all") so it would revert to its default proper value "Invert selection". — AlexSm 19:23, 12 October 2011 (UTC)
Fixed, thanks. Amalthea 19:29, 12 October 2011 (UTC)

Annoying changes

Two recent changes which are annoying the crap out of me:

  1. The search box's AJAX load of matching article titles doesn't seem to work consistently. Sometimes the matches load, but mostly they don't. Also, why is that when I click the article I want, I don't automatically go there anymore? Now all it does is fill the article title in the box and I still have to click Go/press enter.
  2. The cite button in the toolbar. Cite book and web don't look up the information anymore. Instead, it submits the form and I end up saving an edit I was only halfway done with.

Anyone know what's up with these? Thanks. howcheng {chat} 18:17, 12 October 2011 (UTC)

Re #1, I was about to post the same issue myself... the new search box behavior (at least under MonoBook) is very frustrating. In addition to Howcheng's comments, it seems the matches are now constrained horizontally to the same width as the search box instead of allowed to expand to the width of the matches, which really hurts usability. 28bytes (talk) 18:24, 12 October 2011 (UTC)
On Firefox 6 on Windows 7 I have no problems with automatically getting to a particular page but I never go to pages with long titles.Jasper Deng (talk) 18:54, 12 October 2011 (UTC)
I had noticed the same issues with search. Monobook, FF7, Win7. ---— Gadget850 (Ed) talk 20:12, 12 October 2011 (UTC)
Re #1: Reported: Selecting a suggestion in monobook doesn't go to the article
Re #2: I don't know much about the cite button, but that seems like the issue described in this bugzilla report that User:TheDJ said he would attempt to fix. — MarkAHershberger(talk) 21:00, 12 October 2011 (UTC)

Header not showing up

On Time zone, header 1.3 ("Worldwide time zones") isn't showing up for me (Firefox 7 on Linux). Is that because it's below a quote box template and something is malfunctioning? -- Beland (talk) 18:44, 12 October 2011 (UTC)

A {{clear}} just before the header fixes it. I haven't saved the edit, though, in case anyone else wants to look at the problem. -- John of Reading (talk) 18:52, 12 October 2011 (UTC)
The header is actually there, it's just hidden behind the {{quote box}} - this may be demonstrated by editing the whole page, and altering the line:
{{Quote box|quote=The question of uniform time standards ...
to:
{{Quote box|width=95%|quote=The question of uniform time standards ...
and previewing the page: the "lost" heading is visible to the left of that quote box, although truncated. This effect may be a result of putting such a large quote inside a template intended for right-alignment. Anyway, I've applied a better fix than {{clear}} by centring the quote box, see this edit. --Redrose64 (talk) 19:45, 12 October 2011 (UTC)

Hyphen to n-dash

hello,

can you replace the hyphen with an en-dashes on pages? For example, if you have a browser with tabs, it display the title "Wikipedia:Help desk - Wikipedia, the free..." instead of "Wikipedia:Help desk – Wikipedia, the free...". Yes it is nitpicky, but it looks better :P. Thanks. --♫GoP♫TCN 18:55, 12 October 2011 (UTC)

I doubt it - the data in the tab comes from the page header and is the first line of the HTML header and appears as <title>Wikipedia:Help desk - Wikipedia, the free encyclopedia</title> if you view the page source. So that would suggest that it's the Wiki software adding the " - Wikipedia, the free encyclopedia" to the page name when the html page is generated and sent to your browser.  Ronhjones  (Talk) 19:04, 12 October 2011 (UTC)
It can be changed through MediaWiki:Pagetitle (it has been changed the way you suggest before), but be aware that a dash could look funny in some older browsers. Titoxd(?!? - cool stuff) 19:28, 12 October 2011 (UTC)
One learns something here every day...  Ronhjones  (Talk) 19:58, 12 October 2011 (UTC)

Please tell your story for the Wikimedia 2011 fundraiser

When I first signed in I was told I should tell my story. I did click on the link once and was sent to Wikimedia, where I was asked to say why I contributed to Wikipedia and possibly other Wikimedia projects. It seems like every other time I go to a new Wikipedia page this message is at the top, and if I click on the X to get rid of it, I am actually sent back there.Vchimpanzee · talk · contributions · 19:46, 12 October 2011 (UTC)

Update: the last time I clicked on the X to get rid of it, it did go away. Perhaps I didn't aim correctly.Vchimpanzee · talk · contributions · 19:51, 12 October 2011 (UTC)
There is a problem with the extension currently that means sometimes the banner can be coded to not go away properly when you attempt to dismiss it. There is a fix that has not yet been deployed to make it impossible to incorrectly code it in the future. Please continue to report any problems. — MarkAHershberger(talk) 20:00, 12 October 2011 (UTC)
I checked on the coding for that banner and it looks OK. The close box for that banner was actually coded incorrectly when it was first used, but was fixed a few days ago. Perhaps you got a cached version of the banner. Kaldari (talk) 20:23, 12 October 2011 (UTC)

Highlighting problem in editor

Sometimes I have trouble highlighting text, in preparation to copy it. If I left-click & drag from the middle of the edit window toward the bottom, the text will not scroll when I reach the bottom of the window. But if I drag toward the top of the window, the text will scroll so I can highlight all the text I want.

I am using Firefox 7.0.1, Windows 7, and the Vector skin. — Preceding unsigned comment added by Jc3s5h (talkcontribs)

I rarely use mouse dragging to highlight text in the edit window but it works for me in Firefox 7.0.1, Windows Vista, Vector skin. I doubt Wikipedia can do anything to change the behaviour for you. Instead you can place the text cursor at the start or end, hold down shift and move with keys (arrows, page up/down, etc.). PrimeHunter (talk) 01:01, 13 October 2011 (UTC)
I had never bothered to learn the Windows highlight-with-keys method, so thanks for that hint. Jc3s5h (talk) 01:29, 13 October 2011 (UTC)

1.18 issues tagged as resolved con't

Watchlist RSS feed now acts as "expand watchlist..."

  Resolved
 – Fixed in rev:99694.

For some reason, perhaps related to 1.18 rollout, the RSS-feed of my watchlist now acts like "Expand watchlist to show all changes, not just the most recent" - which is undesirable since the RSS feed already only shows a limited number of changes. –xenotalk 18:37, 12 October 2011 (UTC)

Hm - after finding mw:API:Feedwatchlist, I guess I should pass the "allrev" option but I can't seem to figure out what the value should be. "No", "False" and "0" do not seem to work. –xenotalk 19:10, 12 October 2011 (UTC)
It's a bug. The correct answer is "do not specify allrev"; flag-style parameters in the API are true if they are specified at all, even if the value is 0 or the empty string. It appears to have been broken in rev:85756, Reedy changed the default for the "allrev" parameter from null to false but forgot to change line 103 to match. Anomie 20:09, 12 October 2011 (UTC)
Thanks. I've commented at that revision and re-opened bugzilla:11218. –xenotalk 02:04, 13 October 2011 (UTC)
This has since been resolved. Thanks to all for the quick fix. –xenotalk 22:15, 13 October 2011 (UTC)

Refn (#ref) broken

  Resolved
 – Template corrected
Extended content

I have just added some testcases to {{Refn}}, see Template:Refn/testcases. A second unnamed reference generated by Refn does not appear in its Reflist, although a reference embedded within it does appear. According to the documentation, the name is supposed to be optional. Adding a unique name causes the corresponding note to appear. --Mirokado (talk) 19:31, 12 October 2011 (UTC)

Not broken, I have simply procrastinated on updating the documentation. ---— Gadget850 (Ed) talk 19:49, 12 October 2011 (UTC)
Wait— I did update it. Something odd here. ---— Gadget850 (Ed) talk 19:59, 12 October 2011 (UTC)
  Working #tag:ref works, but {{refn}} doesn't work unless you use a name. <references /> and #tag:references work, but {{reflist}} is doing something odd. I have this up for some folks to look at. ---— Gadget850 (Ed) talk 20:35, 12 October 2011 (UTC)

  Fixed Please let me know if there are further issues. I am watching the template page, so I would probably see messages there that might get buried here. ---— Gadget850 (Ed) talk 11:54, 13 October 2011 (UTC)

Thanks for the quick response. testcases look fine now. Anything else can be discussed on the template page as you suggest. --Mirokado (talk) 14:49, 13 October 2011 (UTC)

Category or list of superseded images

Occasionally I'll come across an image in an article that's a JPEG and shouldn't be (a diagram/chart/logo or similar). Generally I tag the file with BadJPEG or ShouldBeSVG, etc. but sometimes the image has already been superseded and has not been replaced in all (or any) articles. Is there a category or list of such images, so I can see all images which have been superseded but not replaced in articles? Forgive me if there already is such a thing and I've just somehow missed it. —danhash (talk) 21:27, 12 October 2011 (UTC)

Template:Vector version available puts images in Category:Wikipedia images available as SVG and Template:PNG version available puts images in Category:Images made obsolete by a PNG version, but those include images which are no longer used. I suppose you could use CatScan to find intersections with Category:Orphan images like so. Not exactly the most elegant of solutions, but I suppose it'll do? —Calvin 1998 (t·c) 01:42, 13 October 2011 (UTC)
Thanks! That's pretty cool. I typed out another comment here and then realized I was an idiot and your CatScan criteria excluded orphans instead of including them–I guess I got confused by your comment. But it works! So thank you :)
I couldn't find an option on CatScan for filtering by if an image is used in a specific namespace, or even used at all. Is there a way to only show images which are used in article space on the English Wikipedia?
Part of this could just be a simple SQL query, but AFAIK an image's use in a certain namespace isn't necessarily a separate, sortable column in the database, so this might involve more complex/intensive processing? I'd settle for an imperfect version, but the less is done by a computer the more I'll have to do by hand.
Also, does CatScan find images on Commons that are used on Wikipedia, or just images uploaded locally to Wikipedia?
Another question (I'm full of them I know): Is there an semi-automated way to change multiple inclusions of an image to point to a different image, similar to what twinkle does with unlink or the way popups can fix links to redirects/disambiguations? It'd obviously need manual operation to avoid changing user pages and pages in other namespaces unnecessarily. It's pretty uncontroversial to supersede images in article space, but I have no need to change them elsewhere.
Thanks!! —danhash (talk) 15:01, 13 October 2011 (UTC)
For an image like this, it has been superseded, and the old version isn't in use any more and somebody already switched the article it was used in to use the new version, but it is not tagged with {{orphan image}}, so therefore not in Category:Orphan images, so still showed up on the CatScan list. On reading the orphan image template, it's pretty specific (only for freely licensed images). There is {{obsolete}}, but that would be somewhat redundant with {{vector version available}} wouldn't it? Any ideas? —danhash (talk) 15:27, 13 October 2011 (UTC)
Also, is there a way to restrict the list to only JPEGs? I'd like to get to all of these images, but most important to me is fixing JPEGs in articles which now have versions without compression artifacts, whether the new version is PNG or SVG. This is a lot of questions at once I know and I realize it may take some time to figure this out entirely. The most important part, though, would be the semi-automation of changing the image references. For example, there are over 100 references to File:Massachusetts_coat_of_arms.png to change to File:Coat_of_arms_of_Massachusetts.svg. Thanks. —danhash (talk) 15:46, 13 October 2011 (UTC)
WP:AWB can probably help with the "find links to a file, replace with a different file". Can't do it completely blindly, since the images might be different sizes (need to add |px parameter?). I would love to know of an alternate tool because AWB is Windows-centric. DMacks (talk) 15:54, 13 October 2011 (UTC)
Hadn't thought of AWB. Is there possibly anything else on the toolserver that could help with this? —danhash (talk) 16:35, 13 October 2011 (UTC)

Technical questions

hello,

can you answer the questions on Wikipedia:Help_desk#Watchlist_question? Maybe I am not the first one who reports those bugs. Thanks.--♫GoP♫TCN 09:44, 13 October 2011 (UTC)

Problems with User:Animum/easyblock.js

  Resolved
 – Fixed in rev:99738. Thanks for quick work, all. –xenotalk 22:32, 13 October 2011 (UTC)

Hi. I've been using User:Animum/easyblock.js for a while. Starting quite recently (within the past few days), it disables talk page access when blocking an editor. It never did that before. Any suggestions?

I'm running Firefox 8.0a in Windows. — Malik Shabazz Talk/Stalk 16:52, 13 October 2011 (UTC)

Likely related to #Changed appearance of Special:Block. –xenotalk 16:54, 13 October 2011 (UTC)
Yes, that's probably it. Thanks for the quick response. — Malik Shabazz Talk/Stalk 17:23, 13 October 2011 (UTC)
Only in that someone forgot a "!" when changing all that around. See T33679. Anomie 20:35, 13 October 2011 (UTC)