Wikipedia:Village pump (technical)

The technical section of the village pump is used to discuss technical issues about Wikipedia. Bugs and feature requests should be made at Bugzilla (How to report a bug). Bugs with security implications should be reported to securitywikimedia.org. Newcomers to the technical village pump are encouraged to read these guidelines prior to posting here. Questions about MediaWiki in general should be posted at the MediaWiki support desk.
Please add new topics to the bottom of this page.
« Older discussions, 89, 90, 91, 92, 93, 94, 95, 96, 97

Centralized discussion
Proposals	Discussions	Recurring proposals
RfC about allowing watchlisting of user contributions Binding community discussion on the Abortion article titles, as requested by the Arbitration Committee. RfC on the proposed cessation of selective delete (other than history merge fixes) and resulting changes to deletion policies, including the removal of a Revision Deletion criterion. RfC on how files from non-copyright states should be treated and used on Wikipedia Note: inactive discussions, closed or not, should be archived.
archive · talk · edit · history · watch

Unified login

Resolved

Not sure if I'm in the right place, but after using {{help me}} I was advised to ask my question here.
When logging into Wikipedia through unified login, I'm logged into every project except Wiktionary ^{(see revision in reply of February 10)} (and of course the three two projects pending SUL-request). At first this was solved by not using the secure server (https://), but now it refuses regardless of settings. When logging into Wiktionary directly there is no problem, but then I won't be logged into any of the other projects. In both cases the icons shown suggest I should be logged into all projects. To make it even more interesting: when logging into Wiktionary (manually) after logging into Wikipedia, then logging out on Wiktionary, all of a sudden I am logged out on Wikipedia too, so for logging out it does seem to work... I haven't been able to find anything in the help sections, does anybody have a suggestion? Thank you, — Quibus (talk) 14:00, 8 February 2012 (UTC)

Curious. When I'm logged into Wikipedia, going for wikt: shows the "Redrose64 / My talk / My preferences / My watchlist / My new messages (None) / My contributions / Log out" links upper right, so I'm logged in there too. If I then log out of Wiktionary, and return to Wikipedia, I find that I have been logged out there too. This is what I expect SUL to do.

I have observed the following over a period of many months, which may not be the whole story. The process of logging out zaps the cookie created when you logged in. I think that SUL means that one cookie is shared by all projects: therefore, logging out of Wiktionary will also log you out of Wikipedia. When additional projects join SUL, the user database is not always copied over, so it's worth visiting Special:MergeAccount every so often to keep them up to date.

When the secure server had its own URL structure, i.e. https://secure.wikimedia.org/wikipedia/en/wiki/Main_Page for the Main Page, it had its own cookies. Now that the only difference in URL structure is https versus http (the secure server main page is now at https://en.wikipedia.org/wiki/Main_Page), it can be demonstrated that by logging in "insecure" and then switching to "secure" (by altering http to https in the browser address bar) you are still logged in, therefore the cookie is shared. --Redrose64 (talk) 15:38, 8 February 2012 (UTC)

The cookies are only shared if you log in insecurely. If you log in using https, the cookies are marked "secure" so you will not be logged in when accessing via http. Anomie⚔ 17:30, 8 February 2012 (UTC)

Thank you both for your input and suggestions. I have visited Special:MergeAccount every so often since my SUL-request was initiated. After reading your posts I've checked whether removing cookies would do the trick (it didn't) and if logging in through secure server again would change anything (it also didn't). When checking the projects in the Special:MergeAccount list on whether I'm logged in or not I noticed Wikitionary is not the only project I'm not logged in, it's every project under Wikimedia I'm not logged into and a couple of foreign Wikis:

• Unified login succesful: af, en, es, fo, fr, ga*, gd*,it, nl, pl, ru, tpi, uk
• Unsuccesful: Wikicommons, Wikibooks, Wikinews, Wikiquote, Wikisource, Wikiversity, Wiktionary, fo.wikisource, ga.wiktionary*, gd.wiktionary*, Wikimedia incubator, Mediawiki, Wikimedia Meta-wiki, Wikispecies

So, issue not solved, but maybe this makes things clearer? — Quibus (talk) 09:13, 10 February 2012 (UTC) * = modified after PrimeHunter post, — Quibus (talk) 16:16, 10 February 2012 (UTC)

You have accounts at gawiktionary and gdwiktionary but not gawiki and gdwiki, so ga and gd in your Unsuccesful must be at wiktionary. Then it appears Wikipedia is treated specially for you. Some browsers can make cookie settings for individual domains. Maybe you accidentally did this for wikipedia.org. Which browser do you have? Can you try another? PrimeHunter (talk) 12:57, 10 February 2012 (UTC)

Thank you PrimeHunter, you're absolutely right about the ga and gd account being Wiktionary, I didn't notice. After visiting them again (still not logged in) I also visited ga-wiki and gd-wiki, where the unified login did work, so they both appear in the Special:MergeAccount list now.

I didn't think of using another browser, thanks for that one, sadly it doesn't make a difference. Standard browser is Firefox 9, alternative is IE9. I'll try checking cooky settings now, in the mean time any suggestion is welcome. — Quibus (talk) 16:16, 10 February 2012 (UTC)

Resolved! Looking at the cooky settings I noticed I had third party cookies disabled. Enabling it did the trick and everything works fine on https:// too. Thanks again everybody for all the input. — Quibus (talk) 19:09, 10 February 2012 (UTC)

I've found that Special:MergeAccount doesn't always pull in all other projects. If you check http://toolserver.org/~luxo/contributions/contributions.php?user=Xxxx (where Xxxx should be replaced by your username) you may sometimes see that one or more of the listed projects has a pink background; near the top it says "SUL: Account unattached" instead of "SUL: Account automatically created ...". Click on any page listed in the pink region, go for your browser's "back" button then the "refresh" button, and you should find that the pink background disappears indicating that your accounts are now linked. --Redrose64 (talk) 14:23, 17 February 2012 (UTC)

This is unrelated to Quibus' problem. What you describe is almost certainly cases where a page you have edited has been transwikied to a wiki where your account is not registered yet. Just like other wikis, your account is automatically registered first time you visit any page there. I admit it's confusing that your edits can be registered in a wiki where your account is not registered. PrimeHunter (talk) 21:37, 18 February 2012 (UTC)

Asian font for a punctuation character?

I asked this at the Help desk and was recommended to ask here: At Ellipsis – in Japanese and Chinese, the 3-dot leaders “…” (U+2026) appear on the baseline when the text displays in a Western font. In Asian typesetting it should be vertically centered, and in fact the same codepoint in Japanese (& some Chinese) fonts displays properly centered (like 3 midpoints ···). Is there a template or something with a CSS rule to force display of an East Asian font? I know Unicode obviates most font assignments for world scripts, but in this case it makes a difference visually. I considered substituting the midline ellipsis math operator (⋯ U+22EF) but that would be misleading about the correct character to use. So I thought I should ask for opinions. MJ (t • c) 19:27, 11 February 2012 (UTC)

You might try template {{lang}}. It won't force a font, but might trigger your browser to pick the right one. — Edokter (talk) — 22:06, 11 February 2012 (UTC)

Thanks, Edokter – {{lang}} sounds like exactly what I wanted. But then I went to edit the section in Ellipsis and found it already used there. I tried changing {lang|ja} to {lang|hani}, but it made no difference. I guess that’s because the ellipsis is in General Punctuation, not an Asian block of Unicode. MJ (t • c) 02:24, 12 February 2012 (UTC)

In my environment (Japnese Windowx XP, IE8), {{lang|ja|…}} displayes properly vertically centered dots (…), while without the template, the dots appear on the baseline (…). It seems the markup is fine and the problem is that the font your browser is using for Japanese is not the one with vertically centered dots for horizontal ellipsis. --Kusunose 06:38, 21 February 2012 (UTC)

HTTPS issues

Browsing https://en.wikipedia.org in Google Chrome, the HTTPS encoding is displayed red (№ 4 in this table). It was green before, so something insecure seems to have been added recently. See also broader discussion on Meta. It Is Me Here ^{t / c} 14:40, 12 February 2012 (UTC)

It works for me, so it suggests that it's one of your user scripts or gadgets that's at fault. I shall have a poke. - Jarry1250 ^{[Deliberation needed]} 14:50, 12 February 2012 (UTC)

Hmm, well I'm running prose_size.js, so it can't be that (I'm assuming you're running vector). What gadgets have you got enabled? - Jarry1250 ^{[Deliberation needed]} 14:52, 12 February 2012 (UTC)

It is green for me.

But there are gadgets which were not migrated to protocol relative URLs yet. E.g. if I enable Wikipedia:Multimedia beta features I get the following (because of bugzilla:29649, I think):

The page at https://en.wikipedia.org/wiki/Main_Page ran insecure content from http://prototype.wikimedia.org/mwe-gadget/mwEmbed/remotes/mediaWiki.js?uselang=en.

There may be some other gadgets with the same problem. Helder 14:54, 12 February 2012 (UTC)

Not really bugzilla:29649, although that may well be an issue for the gadget too. The cause of the problem here is that the gadget loads http://prototype.wikimedia.org/mwe-gadget/mwEmbed/remotes/mediaWiki.js?uselang=en. And since prototype.wikimedia.org doesn't seem to support https access, we can't really do anything to fix it. Anomie⚔ 15:47, 12 February 2012 (UTC)

Vector (css, js); common js. Gadgets: Browsing # 3, 6, 7, 10. Editing # 2, 3, 7. Appearance # 1, 3, 7, 8, 16, 18. Compatibility # 2. Advanced # 2. It Is Me Here ^{t / c} 15:02, 12 February 2012 (UTC)

Appearance #16 would be the Wikipedia:Multimedia beta mentioned above, wouldn't it? Try disabling that one temporarily and see if that fixes your issue. Anomie⚔ 15:47, 12 February 2012 (UTC)

I'm afraid that unticking it in Preferences left the icon red, even after purge/hard-refresh. It Is Me Here ^{t / c} 10:25, 13 February 2012 (UTC)

Since you've made a note of which you've got enabled, perhaps you could try disabling them all temporarily? - Jarry1250 ^{[Deliberation needed]} 13:49, 13 February 2012 (UTC)

I have to report that now that I've unticked MWEmbed the padlock is green again in Chrome (although I could have sworn that it continued to be red initially after removing the gadget from my preferences). It Is Me Here ^{t / c} 13:59, 17 February 2012 (UTC)

The server which is hosting that gadget, unfortunately isn't available on https atm... It's in the works of going trough code review, so i doubt it will be updated before that is finished... I'll ask anyway though. —TheDJ (talk • contribs) 12:09, 18 February 2012 (UTC)

Printable version without images

Is there a way to generate a printable version of an article, that would not include images (and their descriptions)? --eugrus (talk) 19:19, 13 February 2012 (UTC)

This will suppress images. Need to dig a bit to get the captions. Will get back to this in a bit.

/* do not print images or captions*/
@media print {
     img {display: none;}
    .thumb {display: none;}
    .thumbinner {display: none;}
    .thumbcaption {display: none;}
    .magnify {display: none;}
}

---— Gadget850 (Ed) ^talk 19:59, 13 February 2012 (UTC)

I've tried putting into User:Eugrus/common.js and it doesn't effect anything at all. Isn't it the right place for this script? --eugrus (talk) 22:03, 13 February 2012 (UTC)

You want to put it in User:Eugrus/common.css instead. Titoxd^{(?!? - cool stuff)} 08:26, 14 February 2012 (UTC)

Tried this as well. No effect( --eugrus (talk) 10:19, 14 February 2012 (UTC)

I see that you've left the code in User:Eugrus/common.js - you should probably blank that out. Also, make sure that you bypassed your browser's cache, as described at the top of the .css and .js pages. --Redrose64 (talk) 11:22, 14 February 2012 (UTC)

I left that a bit vague didn't I. Got a real life interrupt while I was writing that. I updated the source code to suppress captions. ---— Gadget850 (Ed) ^talk 11:48, 14 February 2012 (UTC)

/* do not print images or captions*/
@media print {
     img,
    .thumb,
    .thumbinner,
    .thumbcaption,
    .magnify {display: none;}
}

Works also and requires less css. Not that it matters because it ends up being minified (to some extent; mostly the whitespace), but just noting. --Izno (talk) 19:53, 14 February 2012 (UTC)

I am now getting where the misunderstanding is: what I am talking about is getting rid of images on "&printable=yes" pages: not just in the actually printed text on paper. Is this possible? --eugrus (talk) 22:20, 15 February 2012 (UTC)

My gadget Print dialog might be useful. —TheDJ (talk • contribs) 12:13, 18 February 2012 (UTC)

Random Sampling of AFDs?

Have a look at Template:Recent changes article requests. I'd like to do the same sort of thing, except transcluding the output onto a single page and generating the list of possible titles from the contents of Category:AfD debates. Is there a way that a template could parse that category list and pick random entries, or would we need a bot to generate such a list (in the form of Template:Recent changes article requests/list, for example)? The idea is similar to that of the "Random Article" link, in that you click and get a set of debates you might not ordinarily see. Perhaps articles from outside your usual areas of interest, or in categories you don't normally review. The age of the debates is a factor as well; unless it's the first or last day of the debate, it's difficult to highlight that the article is still up for deletion (apart from the article itself). So this might get more eyes on articles in the middle few days of their debates. Is this feasible? Or am I overthinking it? UltraExactZZ ^Said ~ Did 19:46, 13 February 2012 (UTC)

As far as I can tell, it wouldn't be possible to fully automate this type of process using only templates. You can get a list of articles at AfD by using something like {{#categorytree:AfD debates|mode=pages|hideroot}}, but it would be formatted into HTML already, and furthermore there isn't a practical way (that I know of) to just pick a few out of the list randomly and transclude them to a page.

However, this would be a trivial task for a bot to do, and as you indicated on my talk page, the bot I run already works with AfD's and is fully aware of the status of all articles that are currently at AfD. I think this is an idea worthy of a bot task, after some discussion and fleshing out of the details. There are probably several different kinds of such lists that could be created simultaneously. In addition to a list of random open AfD's that is updated a few times per day (perhaps about 25 AfD's on the list), I think it would be useful to maintain a list of AfD's that urgently need more attention from potential voters. The bot could assign an "urgency score" to each AfD and then list the top 25 AfD's by score. I'm envisioning the score being comprised of things like:

1. The fewer bolded (non-comment) votes an AfD has, the higher its score
2. The smaller the AfD discussion page is (i.e. in bytes), the higher its score
3. The shorter the time before the AfD is scheduled to be closed, the higher its score
4. The more times an AfD has already been relisted, the higher its score

Any comments, suggested tweaks, or ideas for other lists? —SW— ^babble 20:54, 13 February 2012 (UTC)

Much simpler than my idea - I like it. You could also check for stagnant debates, using the time since last edit measurement. The trick is coming up with a metric for everything rolled into one score. UltraExactZZ ^Said ~ Did 13:18, 14 February 2012 (UTC)

I've got something preliminary set up at User:Snotbot/AfD's requiring attention. —SW— ^converse 00:03, 15 February 2012 (UTC)

I like it. It's self-evident why the debates are listed (most for lack of participation). Does the bot create a report somewhere that shows actual scores? UltraExactZZ ^Said ~ Did 19:17, 16 February 2012 (UTC)

The bot seems to be gravitating toward the most relisted debates, which probably comes from the scoring criteria - so that's working. I think this'll do just fine. Thanks! UltraExactZZ ^Said ~ Did 13:26, 21 February 2012 (UTC)

"Forgot my password" from email

I've proposed that the new feature to email generated temporary passwords based on the email address be enabled. Superm401 - Talk 06:33, 14 February 2012 (UTC)

I hope that WP does not keep any user passwords! This would be most insecure. See password hash and the sony debacle : PlayStation_Network_outage#Unencrypted_personal_details 21:22, 16 February 2012 (UTC) — Preceding unsigned comment added by 129.67.86.189 (talk)

Looking at bugzilla:34386 and bugzilla:13015, I think the new feature will actually send an email containing a forgotten account name. There's nothing there about sending the password. -- John of Reading (talk) 21:28, 16 February 2012 (UTC)

Em 34386| talks about passwords and then changed to username half way through. And WMF would already have details of passwords stored somewhere or you wouldn't be able to log in.Edinburgh Wanderer 21:31, 16 February 2012 (UTC)

No, they don't store passwords - they store encryptions of passwords. During the logging-in process, when you enter a password, it's encrypted and this encrypted version is compared against the encrypted password stored in the database. If these match, the password is assumed to match as well: the chances of two different passwords having the same encrypted form are extremely small. There is no actual means of decrypting the encrypted form. --Redrose64 (talk) 22:31, 16 February 2012 (UTC)

To be pedantic, "encrypted" is the wrong word there. "Hashed" is the correct term. Anomie⚔ 02:00, 17 February 2012 (UTC)

Thats not true - a password hash is one way only --you do *not* need a copy of a password to log in. You need the salted, hashed password. 86.7.36.50 (talk) 22:25, 16 February 2012 (UTC)

That's technically not true either. You need something (not necessarily the original password) that, after the authentication code salts and hashes it, matches the stored password hash. Anomie⚔ 02:00, 17 February 2012 (UTC)

I'm sorry for the unclear wording. I've clarified it. Wikipedia certainly does not store user passwords in plain text. It uses per-user salted double MD5. It currently has a feature to email you your username and a new temporary password if you forget your current password. You also don't have to use the temporary password, in case it is sent maliciously; it expires in 7 days. The proposed feature is to email this information even if you forget your username. Superm401 - Talk 00:22, 21 February 2012 (UTC)

Suggested edit to template and creation of tracking category

{{Rating}} This template accepts values such as {{Rating|9|10}} but not {{Rating|9.1|10}} (as it appears on this revision of Paranoid (album).) Can someone please amend this so that incorrect uses populate a tracking category? I'd be happy to sweep through and fix them. Thanks. Please also respond on my talk if you can to let me know that it's been made. —Justin (koavf)❤T☮C☺M☯ 07:40, 15 February 2012 (UTC)

From the Template Documentation: For example, if a reviewer awards something a rating of "8.6" (expressed in digits) on a scale of ten, you should write it as "8.6" or "8.6/10.0" in your article. Do not use {{Rating|8.6|10}} as it is inaccurate and misleading. (Emphasis is mine. ) It's not a technical issue. Setting it up so that it can show a fraction of a star is technically complex, and probably not worth it. PuppyOnTheRadio (talk) 08:16, 15 February 2012 (UTC)

Koavf isn't asking for the ability to display fractional stars. He's asking for a means to detect such attempts at misuse, I suspect that it would be somewhat along the lines of Category:Pages with incorrect use of RailGauge template which lists attempts to use an unrecognised value in the {{RailGauge}} template. For example, {{RailGauge|1435}} is valid, but {{RailGauge|1430}} is not, and will put the article into that tracking cat. --Redrose64 (talk) 14:06, 15 February 2012 (UTC)

Indeed. But the documentation does not actually state that fractional numbers should not be used, it states only that the template should not be used for ratings that were not originally expressed in "stars". The template code seems to round any fraction to a half, 1.01, 1.5 and 1.99 would all display identically, like this: , , . Before changing the template, which is edit-protected, perhaps someone familiar with its use could edit its unprotected documentation page with more restrictive guidance, to demonstrate that there is a consensus for prohibiting some (or all?) decimals. Koav might not have realised that the template accepts decimals and displays halves. — Richardguk (talk) 15:13, 15 February 2012 (UTC)

This is a non-issue I think. Reviewers that use stars don't use tenths. The rating in the example should just say "9.1" and shouldn't be using {{rating}} at all (it's actually incorrect now, as the mouseover says 9/10 instead of 9.1/10). As for the tracking category, no opinion, although hopefully it won't be used to round down decimals like on Paranoid. — Bility (talk) 18:44, 15 February 2012 (UTC)

Clarification RedRose is exactly right--I'd like a tracking category like the example given, for exactly the reason that Bility points out: no one assigns "9.1 stars" so any time that appears in an article 1.) the template has been misused and 2.) the template incorrectly rounds these values. It should be amended so that it only accepts "X", "X.5", and "X.0" values and if anyone puts in "X.7" then an error tracking category is generated and someone (me) can come along to fix these errors. —Justin (koavf)❤T☮C☺M☯ 19:14, 15 February 2012 (UTC)

Category I've done the easy part: Category:Pages with incorrect use of Rating template. —Justin (koavf)❤T☮C☺M☯ 21:58, 15 February 2012 (UTC)

In which case, my couple of edits to the page in question could happily be undone. I might just do that. PuppyOnTheRadio (talk) 00:39, 16 February 2012 (UTC)

I have limited access to the site at work, so my ability to test is a little limited, but as far as can work out all that it needs is:

<includeonly>{{#ifexpr:{{{1}}} mod 0.5 = 0|<!-- There is no issue.  Nothing to see here.  Move along -->|[[Category:Pages with incorrect use of Rating template]]}}</includeonly>

to be popped in there. PuppyOnTheRadio (talk)

Another alternative is to not worry too much about the categorisation, but instead set the template up so that if it picks up the catch I've set up there it dislays it as a numerical figure instead, and otherwise as per the current template set-up. PuppyOnTheRadio (talk) 01:01, 16 February 2012 (UTC)

It's already implemented in Template:Rating/sandbox, although unfortunately the wiki version of the modulus operator isn't refined enough to use decimals. There are other changes in the sandbox, but I did mention this tracking category in the talk page thread. — Bility (talk) 01:03, 16 February 2012 (UTC)

Really? I could have sworn I've done that in the past with a binary calculator. Stupid parser. An alternative would be {{#ifexpr: {{{1}}}*2 <> floor({{{1}}}*2)..., but either way works. PuppyOnTheRadio (talk) 01:21, 16 February 2012 (UTC)

I don't think that would work for n.5? I ended up using {{#ifexpr:{{{1}}}-floor{{{1}}}<>0and{{{1}}}-floor{{{1}}}<>.5. — Bility (talk) 01:29, 16 February 2012 (UTC)

Of course it would. 0.5 multiplied by 2 = 1. Any other decimal multiplied by 2 would would return another decimal figure. PuppyOnTheRadio (talk) 07:23, 16 February 2012 (UTC)

Well in your example of 0.5 (half a star), ​0.5 * 2 <> floor(0.5) * 2​ would return true, meaning it would get categorized as being improperly used even though half stars are valid. At any rate, I've tested the Rating/sandbox version in article space and it correctly adds the category or doesn't as it's written. — Bility (talk) 16:14, 16 February 2012 (UTC)

I left out the end bracket. Should be ​0.5 * 2 <> floor(0.5 * 2)​. (I edited the above bit as well.) PuppyOnTheRadio (talk) 00:57, 17 February 2012 (UTC)

Resolved thanks to Tra. PuppyOnTheRadio talk 07:47, 20 February 2012 (UTC)

Check out this 4.3 star review. :P — Bility (talk) 22:45, 20 February 2012 (UTC)

How does one go about getting a new filetype added to the list of allowable uploads?

The title pretty much says it. I don't want to spill the details as we are still working on a model, but its an XML based filetype (aka text based) that could provide an untold benefit to the project. Who deals with this? I know there will need to be community consensus to implement it, but when that has been achieved where does it go? Bugzilla? - ʄɭoʏɗiaɲ ^τ _¢ 17:33, 15 February 2012 (UTC)

Yes, Bugzilla. Do you mean creating an entirely new filetype (as opposed to adding an existing filetype to the list of allowed filetypes)? Ucucha (talk) 21:04, 15 February 2012 (UTC)

No I mean adding a new type (.kml) to the list of allowed filetypes that can be uploaded. - ʄɭoʏɗiaɲ ^τ _¢ 03:44, 16 February 2012 (UTC)

Unfortunately, kml and kmz are a bit problematic bugzilla:26059. —TheDJ (talk • contribs) 12:18, 18 February 2012 (UTC)

"My Sandbox" link

When did this link get added, and where is the deciding discussion? —danhash (talk) 19:25, 15 February 2012 (UTC)

And how can I get rid of {{user sandbox}} at the top and why don't the sections have links? Or is it just me? — Bility (talk) 19:29, 15 February 2012 (UTC)

Uncheck it in gadgets, section "appearance". — AlexSm 19:38, 15 February 2012 (UTC)

WP:Gadget/proposals#My Sandbox. — AlexSm 19:38, 15 February 2012 (UTC)

Thanks Alex; knew I'd seen it proposed somewhere recently. Are interface changes such as this one documented anywhere that's easy to find? When searching for documentation about the change I came up empty. —danhash (talk) 20:59, 15 February 2012 (UTC)

As far as I know, there is no formalized way in which interface changes as made by the English Wikipedia community are published centrally, though usually, this very spot is a good place to keep an eye on. —TheDJ (talk • contribs) 12:21, 18 February 2012 (UTC)

Sandbox, not sandbox

Has the User namespace always treated the case of the first letter of subpages as significant? (I ask because the My Sandbox link goes to Special:MyPage/sandbox instead of Special:MyPage/Sandbox.) Mark Hurd (talk) 23:29, 15 February 2012 (UTC)

Yes— subpages in all allowed namespaces are case sensitive. ---— Gadget850 (Ed) ^talk 23:36, 15 February 2012 (UTC)

#REDIRECT redirects to an older version if not logged in

I've added a paragraph on RGBA LEDs to LED lamp#Technology overview. It's the third one in this section and it starts with "The color rendering of RGB LEDs, however, is worse than one would expect".

This paragraph doesn't show up if I'm (a) not logged in and (b) am going through a redirect, such as LED bulb. Purging helps only until the next time I clear my browser cache.

I see this on two different machines, one on Win7+Firefox3.6.26 and one on WinXP+Firefox3.6.26, and using two different i'net providers, too.

Have I found a MediaWiki bug? --Mkratz (talk) 20:09, 15 February 2012 (UTC)

This is a known problem, see Wikipedia:Village pump (technical)/Archive 96#Parser cache not invalidated for redirect pages. --Redrose64 (talk) 20:34, 15 February 2012 (UTC)

I've made one more try to get some ops-side explanations on this, then I think it will be time to update WP:Purge and WP:Redirect, even if it's only to say "we're not sure and no-one is saying". Franamax (talk) 21:25, 15 February 2012 (UTC)

Thanks for your help; I've just confirmed with LED bulb that the workaround is to purge the redirect page itself. Another redirect, Led lamp, still delivers an at least two weeks-old version to IP editors.

Parsing the "What links here" page for redirects and purging them looks like a nice job for a bot until the ops get around to fix this problem. --Mkratz (talk) 10:47, 22 February 2012 (UTC)

Would it be technically possible to add a "purge" link to the "Redirected from" message temporarily? Something like "(Redirected from Foo. Out of date copy? Purge the page)" - where "Purge" is a link that will purge the redirect. -- John of Reading (talk) 11:03, 22 February 2012 (UTC)

secure.wikimedia.org down?

This breaks the HTTPS everywhere addon... I can still use https://en.wikipedia.org/ 129.67.86.189 (talk) 21:18, 16 February 2012 (UTC)

Upgrade your HTTPS Everywhere. Anomie⚔ 21:39, 16 February 2012 (UTC)

https://secure.wikimedia.org was also down for me but it's up again. PrimeHunter (talk) 14:08, 17 February 2012 (UTC)

Proposal to allow the uploading of *.kml files

As a result of an RfC at WikiProject:Highways, a discovery was made recently that allows us to display linear data on Google Maps and Bing Maps with an extraordinary level of detail. Thanks to User:Dschwen, WikiMiniAtlas already supports the new data. In addition to the simple linear features that we have exploited thus far, there exists the potential to label specific points, or display entire networks (think subway systems, or rivers and their tributaries) on the mapping services. Lots of potential, but still lots of groundwork to lay.

To see a few examples of what this can do, check out:

• River Liffey (link in title), a convoluted river in Ireland
• Ontario Highway 401 (link in the title), a 500-mile highway
• Deansgate (link in the Further reading section), a short urban major roadway that begins and ends inexplicably

Note how in River Liffey we can put markers on the various landmarks along the river's course, as well as including separate lines for the tributaries that feed the river.

The main issue faced thus far is manipulating the data, as we cannot upload a *.kml file. Since they are text-based XML files, the contents can be copy-pasted into a subpage; however this presents a significant hurdle to user-friendliness.

Google Earth, NASA WorldWind, Gaia, and most GIS software can open, manipulate, create and save these files, so they are far easier to create than the standard PDF format. However, the files are article specific, so I feel they aren't appropriate for commons in most cases.

So, essentially, this is a request for community consent to add *.kml to the list of allowable filetypes. - ʄɭoʏɗiaɲ ^τ _¢ 02:44, 17 February 2012 (UTC)

• Support—as a supporter of the KML-based solution, I feel that we need to take this additional step before pursuing the concept on a wider range and number of articles. As it stands, we have no way to easily document the source of the KML data, and using subpages makes it more difficult for readers to reuse the data files. Imzadi 1979 → 02:50, 17 February 2012 (UTC)
• Support *.xml variant = low risk + potential benefit. Happy to say yes. PuppyOnTheRadio (talk) 03:19, 17 February 2012 (UTC)
• Comment See bug 26059http://bugzilla.wikimedia.org/show_bug.cgi?id=26059: at the moment, it seems kml is unlikely, but kmz might be possible. Anomie⚔ 03:43, 17 February 2012 (UTC)

  Is there no way to have wiki output/treat them as raw text instead of trying to parse or run the code? - ʄɭoʏɗiaɲ ^τ _¢ 06:35, 17 February 2012 (UTC)

• Support but repeating my observation at Wikipedia talk:WikiProject Geographical coordinates#...and a few points, are you sure this is not already allowed?
  Regarding KML file uploading: there are already about 120 KML files uploaded at Commons some years ago, made accessible from the parent :File: pages there via commons:Template:Overlay (documented further at commons:Commons:Geocoding/Overlay; e.g. File:Etna eruption seen from the International Space Station.jpg which references commons:File:Etna eruption seen from the International Space Station.jpg/overlay.kml (but note the unexpected redlink if the same KML subpage is accessed indirectly with :File: instead of :commons:File:). I don't know whether KML files can be (or already are) uploaded to enwiki, but it seems unlikely that the ability to upload KML files to Commons will have been revoked since the 120 files were added. — Richardguk (talk) 00:02, 17 February 2012 (UTC)
  — Richardguk (talk) 20:50, 17 February 2012 (UTC)

Given that they are subpages, I'd assume they were manually added by copy/pasting the contents of a kml file, similar to what we are doing here at the moment. - ʄɭoʏɗiaɲ ^τ _¢ 22:16, 17 February 2012 (UTC)

I take your point. From the instructions at Geocoding/Overlay, you are right. It hadn't occurred to me that the difference is significant, but I suppose it means that the existing KML "files" are in fact merely "file description subpages" containing KML code (as their wikitext) but which, as far as the servers are concerned, have no file attached. (Afterthought: In some ways, a file description page is more accessible than a proper upload, being editable in place and simple changes viewable as diffs. But I concede that it is a bit of a kludge.) — Richardguk (talk) 23:32, 17 February 2012 (UTC)

• Support. The current way of using subpages is actually fine, too. But if you guys want to upload the data as files on commons, that would be fine, too. It would just be a bit more difficult for me to obtain the data (due to crosssite scripting prohibitions (unless accessin via Special:FilePath works...)). KMZ would require me to write a decompressing proxy on the toolserver. I would certainly prefer raw uncompressed XML/KML data, as that is very easy to process in JavaScript using XMLHTTPRequests :-). --Dschwen 23:37, 17 February 2012 (UTC)

  Actually, if upload.wikimedia.org serves the file with a Content-encoding: gzip header I wouldn't have to worry about unzipping it, the browser would do it for me. --Dschwen 00:02, 18 February 2012 (UTC)

Graphic for portal link to Academy Award portal isn't loading

When the following is added to articles to link to the Academy Award portal, the graphic isn't loading, and the generic portal image appears:

Is there a fix for this? The image that is intended to be displayed is "Oscar.svg", located at Template:Portal/Images/Academy Award. Northamerica1000^(talk) 18:12, 17 February 2012 (UTC)

See Template:Portal/doc#Image. I think you need to move it to Template:Portal/Images/Academy award, lowercase "a". -- John of Reading (talk) 18:24, 17 February 2012 (UTC)

(edit conflict) Er, File:Oscar.svg is a red and yellow flag. Not sure why that's relevant. Anyway, the problem is a misnamed page. It needs to be at Template:Portal/Images/Academy award. --Redrose64 (talk) 18:26, 17 February 2012 (UTC)

Fixed – After creating the new page suggested above, it works! Added image: Video-x-generic.svg Northamerica1000^(talk) 18:34, 17 February 2012 (UTC)

Looking for a tool

For years now, I've been looking for a tool which would be immensely helpful in matching suspected sockpuppets to possible puppet masters. What I need is a tool into which you can put a number of articles, and it would return the names of editors who had worked on all those articles, or a significant subset of them -- this on the theory that socks frequently return to the "scene of the crime" and can't resist editing the articles they edited under previous IDs. The list generated would provide some starting points to look for behavioral commonalities.

Does anyone know of such a tool, or would anyone be interested in making it? Beyond My Ken (talk) 21:55, 17 February 2012 (UTC)

Pretty easy for someone with a toolserver account. — Bility (talk) 22:21, 17 February 2012 (UTC)

You may be aware of this already, but Wikistalk performs a broadly similar function and can be useful for sock hunting. Graham87 03:53, 18 February 2012 (UTC)

Wikistalk (which is a great tool) actually does the opposite of what I want. It takes multiple editors and tells you what articles they have in common. What I need is a tool which you give articles and it generates users who edited them in common, Beyond My Ken (talk) 05:04, 18 February 2012 (UTC)

You'll want to filter out a lot of bots and other prolific editors, but otherwise, yeah, it sounds straightforward. Got any particular articles in mind? 67.117.145.9 (talk) 00:40, 19 February 2012 (UTC)

Not at this time - I'm lookng for a generally usable tool. Beyond My Ken (talk) 22:59, 19 February 2012 (UTC)

Looking for Help Gathering Data on Bots

Hello. Some of you might have run into me before...I am doing a research project on bots, bot operators, and technical tools on WP and WM projects. I'm wondering if anyone wants to tackle this problem, which would help me out tremendously. I am looking for stats and data on bots, especially over time. Things like:

• (#) of bot accounts registered over time (by month would be fine) (on English WP)
• (#) of bot edits over time (on English WP)
• (#) of BRFA approved and not approved over time (on English WP only, obviously)
• same trends for bot use on other language versions (which would be a bonus)

I've found some info on these things spread around WP, but nothing that is both up-to-date and reasonably accurate/reliable. I'm not sure if getting this info involves dealing with a data dump (I suspect it does), or if there are simpler ways to do it. If you're interested in investigating this with me, I'd really appreciate the help. Please let me know here or on my talk page.

And if you're a bot operator or Wikimedia developer (or someone who deals with the technical infrastructure of WP) and you'd like to be interviewed, please see my call here.

Thanks! UOJComm (talk) 23:42, 17 February 2012 (UTC)

The folks over at Wikipedia:Database reports should be able to help you out on most counts. Josh Parris 05:15, 18 February 2012 (UTC)

It's very hard to gather meaningful data on bot editing of wikipedia since there are an unknown number of unauthorized ones running from regular user accounts at any time. We just had a messy arb case to get rid of an especially persistent abuser, but there are are constantly new incidents. It's like trying to get statistics on liquor consumption by looking at liquor tax receipts, in a region full of moonshiners. 67.117.145.9 (talk) 21:47, 18 February 2012 (UTC)

Thanks to both of you for the help. Yes, I don't imagine I will be able to get a fully accurate take on bot activity because of those flying under the radar (or the ones who apparently think they are...could you pass along the link to that ArbCom case?). For my project, just getting data on approved bots will be good enough, though. I will check with the people at database reports. Thanks again! UOJComm (talk) 00:35, 19 February 2012 (UTC)

The arb case was WP:Arbitration/Requests/Case/Betacommand 3. That situation wasn't "under the radar", it was blatant, it just took years of drama to make it (probably temporarily) stop. One issue is there's a bogus "legal" definition of a bot (WP:BOTPOL) that enables a lot of wikilawyering over whether something is a bot or not, with lots of programs and editing just skirting the boundaries and avoiding getting treated as bots (even though that is what they are). So the arb case I linked didn't involve long-running, fully-automatic bots (that might be what you're interested in) but a lot of smaller bursts of automatic editing. 67.117.145.9 (talk) 00:53, 19 February 2012 (UTC)

What would you suggest to improve the definition? Feel free to start a discussion at WT:BOTPOL. Anomie⚔ 20:54, 19 February 2012 (UTC)

Thank you for the link and the comments. If you would like to be interviewed for my project, I'd love to hear your detailed thoughts. Please send me an email (anonymity can be maintained if desired). UOJComm (talk) 22:49, 20 February 2012 (UTC)

Quartic function has spam inserted directing all clicks to a racist 9/11 conspiracy site

At least for me, going to the page quartic function leads to a strange situation where all clicks on any link (including "edit", "talk", etc.) lead to a racist Jew-baiting 9/11-conspiracy site at pump.pp4l.me. Conceivably this is a virus on my machine but I suspect it's code inserted into the page itself. I can't edit the page or look at its history, but I can see the page source, and I see the following:

<div style="position:fixed;left:0;top:0"><b><a href="http://pump.pp4l.me/proof.php"><img alt="Chess tile .png" src="//upload.wikimedia.org/wikipedia/commons/9/97/Chess_tile_.png" width="8000" height="8000" /></a></b></div>

This is stuck in directly between "This suggests using a ..." and "... resolvent cubic whose roots may be variously described as a discrete Fourier transform or a Hadamard matrix transform ..."

(This is one of the things that suggests a page mod rather than a virus -- it occurs only on one page, and it makes use of Wikipedia-specific links, and it inserts in the middle of some text)

Note that the div places an invisible image over the entire window, intercepting all links. Please fix this and permanently ban whoever put this in!

Thanks.

Benwing (talk) 08:21, 18 February 2012 (UTC)

The vandalized template has already been fixed. Ravensfire (talk) 08:32, 18 February 2012 (UTC)

Yes, see [1]. Goodvac (talk) 08:34, 18 February 2012 (UTC)

I've meta bl'd the link. --Dirk Beetstra ^{T C} 08:36, 18 February 2012 (UTC)

Thanks. Ravensfire (talk) 08:40, 18 February 2012 (UTC)

That was a quite devious abuse of the template mechanism. The "link" mechanism used there looks quite dangerous: is there ever a valid use for it other than to link to other pages on Wikimedia projects? If not, perhaps it should be filtered in software at render time. -- The Anome (talk) 11:05, 18 February 2012 (UTC)

I remember being cought by filter 139 when using style="position: absolute" code, the filter being there to prevent a similar trick. -DePiep (talk) 11:17, 18 February 2012 (UTC)

Filter 139 is currently disabled, apparently because of too many false positives -- otherwise, it looks likely it would have caught this. Perhaps we should have a variant of filter 139 that just checks edits in template space? -- The Anome (talk) 11:27, 18 February 2012 (UTC)

Or which looks for four digit widths and/or heights? Or with high z-index? Or all three? - Jarry1250 ^{[Deliberation needed]} 11:39, 18 February 2012 (UTC)

The really bad thing here was to be able to make images (in this case, an invisible image) link to external sites. A whitelist for the linking mechanism in question would have defeated its principal purpose of redirecting the user to an attack page. I wonder, is there ever any legitimate reason to use this feature to link anywhere but to the same wiki the page itself is on? If not, we could restrict the attribute to use only relative links. -- The Anome (talk) 11:44, 18 February 2012 (UTC)

(edit conflict)When I talked about my edit there, in 2010, the filter was changed to allow established users pass through. I have not followed this issue after that. Filter usage is sensitive wrt the workload, very often there is a tradeoff. -DePiep (talk) 11:47, 18 February 2012 (UTC)

Let's make Filter 139 specific to the template space, then, and re-enable it. Perhaps we should even make the exception admin-only, rather than autoconfirmed? However, I'd still like to see action on making the "link=" parameter safe, as that is a long-term and complete fix to the root problem of this exploit. -- The Anome (talk) 11:53, 18 February 2012 (UTC)

The link parameter is likely to be used in template space. So perhaps a filter that looks for the combination 'link=' and 'position:absolute' would be more effective. — Edokter (talk) — 12:16, 18 February 2012 (UTC)

That is also used in templates. One thing that is used in vandalism such as this is an image linking to an external site - does this have any legitimate uses? Peter E. James (talk) 21:08, 19 February 2012 (UTC)

Add position:fixed as was used in this current abuse. Is the third aspect, the option "transparent image covering whole page" prevented this way? -DePiep (talk) 12:41, 18 February 2012 (UTC)

┌────────────────────────────────────────────────────────────────────────────────────────────────────┘ Okay, they hit a new template with a different but similar link. Anything from the above discussion viable to prevent future attempts? Ravensfire (talk) 17:24, 18 February 2012 (UTC)

Note: We've been getting a lot of complaints about this at the Help Desk. See the following threds:

• WP:HD#Possibly malicious link in page about Royal Thai Police.
• WP:HD#Baldiwn of Forde article may have been vandalized
• WP:HD#Mass vandalism through editing unprotected templates
• WP:HD#All links on OS X page linking to 9/11 site.
• WP:HD#The page of Autism appears to have been hacked with an overlay that provides a link to a racist 9/11 conspircy site
• WP:HD#Anti-Israel Redirects?
• WP:HD#Problem with map on Bosnia and Herzegovina

Singularity42 (talk) 18:15, 18 February 2012 (UTC)

On a related note, we are still getting questions about this at the Help Desk as people need to clear their cache in order for the updated, non-vandalised pages to appear. Would it be possible to temporarily put something in place that either did this for people or told people to refresh their browsers to deal with this? I'm not incredibly well-informed, so I don't know exactly what would be needed, but something would be helpful. ItsZippy ^{(talk • contributions)} 19:13, 18 February 2012 (UTC)

I've now adjusted filter 139 to catch at least some of these, and enabled it. Please could someone check my edits to the filter for sanity, and check that they catch the edits in question? -- The Anome (talk) 19:23, 18 February 2012 (UTC)

I have not checked the HD links.

imo, there are three orifices of attack:

OR1: Cover the whole page with an invisible, linked image

OR2: Link to an external page

OR3: Use style="position:fixed" or ..."position:absolute"

Then there are angles of prevention:

ANG1: Use a filter to prevent some code (like filter 139; could use user-status)

ANG2: Restrict certain code to template space

ANG3: Prevent external links in these templates

By Venn, using multiple lines could prevent many multiple attacks. -DePiep (talk) 19:45, 18 February 2012 (UTC)

422 deals with this as well, stemming from a similar set of attacks last month, though apparently not quite comprehensively; I've emailed the author of that to take a look at it. Shimgray | talk | 19:44, 18 February 2012 (UTC)

I'm glad I found this thread (after stumbling upon a link at ANI). Was trying to figure out where to go to, to point out this weird phenomena. I was reading the London 2012 Olympics article and noticed this strange redirect to a 9/11 Jewish conspiracy page, whenever I clicked on a link, image, or even just the plain white background of the article. I'm highly concerned, as the site it directs me to is causing my anti-virus programme to go berserk, telling me that my computer is under threat. Has anyone else's anti-virus software been flashing up with warnings from this weird site? Wesley☀Mouse 20:00, 18 February 2012 (UTC)

Thanks for adjusting the abuse filter to stop / slow this down. For those that find a page affected by this but aren't sure were it's from, edit the page and look at the templates used in that page. Look at the history for each template that's not protected and you should be able to figure out which one was changed fairly quickly and revert it. Ravensfire (talk) 20:16, 18 February 2012 (UTC)

I've just created and enabled filter 453, which should temporarily stop all recently-created editors from editing templates at all. I'll disable it in a few hours, when this has blown over, and all the affected templates tracked down and fixed, but I think it will be a good idea to keep it around for the time being so that it can be re-enabled at a moment's notice if this starts up again. -- The Anome (talk) 20:37, 18 February 2012 (UTC)

I can't see the filters. It will be all right. -DePiep (talk) 20:51, 18 February 2012 (UTC)

OK: this all seems to have gone quiet for now, and I can't see any more reports of similar vandalism recently. Filter 453 is now deactivated, and filter 139 will stay active for now to watch out for similar attacks. If this recurs, please let me and WP:AN/I know, and further action will be taken. -- The Anome (talk) 22:28, 18 February 2012 (UTC)

OK. I've now deactivated 139, and reactivated 453, which is now hopefully both simple and specific enough to work. The oversighting of the abusive diffs, and the difficulty of extracting diff content without getting an eyeful of hate/malware content, has made this very difficult to debug. -- The Anome (talk) 01:34, 19 February 2012 (UTC)

There is an option in the Misc tab of My Preferences for "Do not show page content below diffs", although I don't know if it works for deleted revisions. It also seems strange that the diffs would be oversighted (and not just deleted via WP:REVDEL), as some of the links are still visible via the spam blacklist. Peter E. James (talk) 21:08, 19 February 2012 (UTC)

That option in My Preferences does indeed work for deleted revisions. Graham87 00:46, 20 February 2012 (UTC)

Adding RSS Feeds to Wikipedia

Is it possible to add RSS feeds of external pages to Wikipedia? If so, how do I do it? For example, I would like an updated feed of my favorite CDs from Listal. I know the <rss> tag works on Intellipedia, but what will work here? Thank you for your help. Allen (talk) 17:48, 18 February 2012 (UTC)

That requires an extension such as RSS, which is not enabled here. --Izno (talk) 18:57, 18 February 2012 (UTC)

See also WP:ELNO no. 9. --Redrose64 (talk) 18:57, 18 February 2012 (UTC)

So, unfortunately, it looks as if it can't be done. Does anyone know if RSS feeds from Wikipedia pages can be put on Listal? Allen (talk) 21:48, 18 February 2012 (UTC)

Racist vandalism back again

Could someone with the appropriate technical skills take a look at WP:AN/I, and try to find out why, despite my best efforts, filter 139 has been missing more of the incoming racist template vandalism? Please be very careful viewing the diffs given there: they cover the content with an invisible image that clickjacks every link to racist sites which may well also contain malware, even in preview mode. -- The Anome (talk) 00:42, 19 February 2012 (UTC)

139 is now deactivated, and only 453 is active, and has been tested on defanged vandal template wikitext. Let's see if it catches the next lot. -- The Anome (talk) 01:35, 19 February 2012 (UTC)

Please contact me immediately if you find any vandalism with a similar MO. Elockid ^(Talk) 03:25, 19 February 2012 (UTC)

There are ongoing efforts on Meta to blacklist the affected urls globally, as they have shown up across multiple Wikimedia wikis. Snowolf ^{How can I help?} 07:08, 21 February 2012 (UTC)

Ways to cut down on load time for a template

I've created a template at Commons (link) that's meant to inform the reader about a lot of copyright information at once. Unfortunately, because copyright is very complicated, there is a ton of template logic inside. The result is that any page which uses the template takes about 7 seconds to load. Is there a way I might cut down on the load time of this template? I cannot think of one, other than forking the whole project into JavaScript and adding it to the site's vector.js. Magog the Ogre (talk) 19:31, 19 February 2012 (UTC)

Haven't got any suggestions on speed but that testing for y or 1-9 character could be made into a separate template. -- WOSlinker (talk) 19:41, 19 February 2012 (UTC)

{{#switch:{{padleft:|1|{{lc:{{{1}}}}}}}
 |y|1|2|3|4|5|6|7|8|9=1 <!-- first character is "y", or any positive integer -->
 |#default=0
}}

Not a bad idea. Of course, I'm sure you can tell that the majority of the logic is in the region specific templates (commons:Category:PD-in templates), commons:Template:PD-in/logic, and commons:Template:PD-in/logic-and. Magog the Ogre (talk) 20:07, 19 February 2012 (UTC)

Of course, you might want to hold off on the port to JS (if it's required in the end) until Lua support goes live (tentatively expected to be near the end of this year), although tangible case studies may induce faster work on the project). - Jarry1250 ^{[Deliberation needed]} 13:32, 20 February 2012 (UTC)

Minor issue with Special:Upload

I've noted a small issue in the Special:Upload for all logos - the summary field includes {{non-free logo}} at the bottom and the licensing options only include "I don't know..." and non-free logo again; thus most logos now are being uploaded with two copyright templates. Minor glitch, but probably easily fixable. Skier Dude (talk) 02:44, 20 February 2012 (UTC)

I second this. I probably double licensed about 50 county seals in the past week or two before realizing this was why. Not a big deal, but probably fixable, like Skier Dude said. --Andrew (User:90) (talk) 20:40, 20 February 2012 (UTC)

Cannot login from iPhone

I have tried during today to use my iPhone's browser to log in and make some edits. This fails on multiple accounts, regardless of which browser I use (tried Safari, the Wikipedia app, and Grazing, a third-party browser; but I suppose they just differ in the chrome, not in the actual HTTP(s) client).

Firstly, there is no "log in" link in the mobile version of Wikipedia. This may be a conscious decision (even as to dissuade mobile users from attempting to log in because of support issues, perhaps??) but I am mentioning this in order to provide the full picture.

Secondly, if I navigate to the main (non-mobile) site's login page, I get redirected to the mobile version http://en.m.wikipedia.org/w/index.php?title=Special:UserLogin&returnto=Main+Page&returntoquery=useformat%3Dmobile%26mobileaction%3Dview_normal_site, which does not display a login box. The warnings on the login page are visible, but not the user name and password boxes. I worked around this by scrolling down to the full site ("View this page on regular Wikipedia") link, as opposed to the mobile site (http://en.wikipedia.org vs http://en.m.wikipedia.org).

But even then, I cannot log in. I obviously know my user name and password, as I was able to log in from my computer. The error message says "Login error / There was an unexpected error logging in. Please try again. If the problem persists, it may be because you have cookies disabled, and you should check that they are enabled in your browser settings". (It then goes on to explain how to secure my account.)

If you can view logs for troubleshooting, I believe my phone's IP address just now is 178.55.164.227.

Brief update: Cookies was set to "From Visited" in Grazing. Just to check that this isn't the issue, I changed it briefly to "Always", but this did not help.-- era (Talk | History) 11:34, 20 February 2012 (UTC)

Top most 1000 used words

Where is list of 1000 or 2000 top most used words of english wikipedia? Is it possible to do this list if it no.--Kaiyr (talk) 14:09, 20 February 2012 (UTC)

I would imagine it would follow the "most used words in English" list... --Izno (talk) 16:21, 20 February 2012 (UTC)

It actually would be nice to generate a concordance of words used in Wikipedia, as this would indicate whether Wikipedia deviates from normal language usage. However, I suspect it would be loaded up with technical terms such as template parameter instructions, and the words "redirect" and "disambiguation". bd2412 T 16:42, 20 February 2012 (UTC)

A possible topic for research, I'm sure, but not really appropriate for encyclopedia writers. Just sayin'. --Izno (talk) 18:46, 20 February 2012 (UTC)

Sort challenge

In User:Sphilbrick/List of current NCAA Division I women's basketball coaches, the table is set up to be sortable. I have a subheading in the first column so that one can edit a row, without having to edit the whole table. That adds a numbering to the entries, which is fine. However, if I sort on the team column, it sort such that 1 is first, then 10, rather than 2.

I tried adding <span style="display:none;">Air Force Falcons</span> inside the headings, but that didn't seem to help; my guess is that it converts the heading code to numbers, then sorts. I tried adding the span style outside the headings, but then it doesn't render the headings.--SPhilbrick(Talk) 16:05, 20 February 2012 (UTC)

What numbering? I don't see any numbers and it sorts fine for me. I went through all your revisions and never saw any numbers. Which skin are you using? — Bility (talk) 16:13, 20 February 2012 (UTC)

You only see numbers in the first column because you have "Auto-number headings" enabled at Special:Preferences#mw-prefsection-rendering. It's disabled by default. PrimeHunter (talk) 16:16, 20 February 2012 (UTC)

I turned on section numbering and tested some invisible spans in the table and it seems to work for me. It look all right on your end? — Bility (talk) 16:31, 20 February 2012 (UTC)

@PrimeHunter Thanks, I've had autonumbering on for so long, I forgot it wasn't a default.

@Bilty I was trying to imbed the invisible spans in the team entry, as oppose to a new line. I see you added through Bradley, is there an automated way to do this, or is it one at a time? In any event, thanks for identifying the problem.--SPhilbrick(Talk) 16:52, 20 February 2012 (UTC)

@Bilty Oh and to answer the question directly, rather than implicitly, yes, that seemed to work. Alabama A&M now comes before Alabama Crimson , but that looks like an oddity in the original source (extracted from Wikipedia:WikiProject_College_Basketball/Master_Table)--SPhilbrick(Talk) 16:57, 20 February 2012 (UTC)

Is the list going to grow? You might want to make a template for these entries in your userspace and use that to cut down on clutter and make it easier to add teams. — Bility (talk) 17:25, 20 February 2012 (UTC)

No, the list is essentially complete (one or two new teams a year show up, but that's pretty minor), although I need to add the sort key to every row, so wondering if that can be done in an organized way rather than one at a time. When you did the first 20 or so, did you do each row manually, or did you have a better technique?--SPhilbrick(Talk) 22:48, 20 February 2012 (UTC)

Did them manually by copy/pasting "<span style="display:none;"></span>" then highlighted the team name and Ctrl-dragged it into the span. If you want, I can do the list fairly quickly with some regex. Do you want to keep the span or make a template? — Bility (talk) 23:00, 20 February 2012 (UTC)

I also need to add the rest of the headings, plus I need to add a sort key for the name, so they will sort by last name , not first. Let me try dumping it to Excel and trying to fix it myself, first. --SPhilbrick(Talk) 23:31, 20 February 2012 (UTC)

I have used {{sortname}} for this purpose; see e.g. first column of table at User:Redrose64#Done, rows for Daniel Kinnear Clark, George Augustus Nokes, and Robert Absalom Thom. --Redrose64 (talk) 14:06, 21 February 2012 (UTC)

Signatures

Is there some (working) way to use different font faces in signature? Armbrust, B.Ed. ^{Let's talk}_{about my edits?} 21:07, 20 February 2012 (UTC)

Yes. See this example: Example (talk). Reaper Eternal (talk) 21:09, 20 February 2012 (UTC)

I don't see any difference after I removed the font-family part of the code. See: Example (talk). Armbrust, B.Ed. ^{Let's talk}_{about my edits?} 21:13, 20 February 2012 (UTC)

If you see no difference, it's possible that the font (in this case Lucida console) isn't installed on your machine. When this is the case, most browsers will fail gracefully, by displaying a fallback font. The example below <span style="font-family:'Courier new',monospace">Courier new, with monospace as fallback</span> shows how this may be amended: if Courier new isn't installed, it'll try monospace; if that isn't installed either, it'll default to the page style. --Redrose64 (talk) 21:31, 20 February 2012 (UTC)

(edit conflict) Sure, just look further up this page at the technicolor riot. There are two main ways: using <font face="xxx">...</font> or <span style="font-family:xxx;">...</span>, where xxx is the font name. Be careful that use of non-default fonts does not violate WP:ACCESS though. I've spotted the following (colours removed for clarity): arial Calibri Copperplate Gothic Light Courier new, with monospace as fallback papyrus Tahoma Trebuchet MS Trebuchet MS (done another way) --Redrose64 (talk) 21:26, 20 February 2012 (UTC)

Thanks. Armbrust, B.Ed. ^{Let's talk}_{about my edits?} 21:36, 20 February 2012 (UTC)

GeoGroup URL bug

There's an unresolved URL encoding issue with {{GeoGroup}}, which requires some attention, please. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 22:04, 20 February 2012 (UTC)

MiszaBot is down

All MiszaBots have been down for three days now. User:Misza13 has not edited since last May and cannot be reached via email, so if anyone has any other ways to reach him/her, can you please do so? Talk pages and noticeboards are starting to pile up and require manual archiving at this time. --Andrew (User:90) (talk) 22:20, 20 February 2012 (UTC)

Also being discussed at Wikipedia:Bot_owners'_noticeboard#Misza13.27s_bots_seem_to_be_down.Edinburgh Wanderer 22:26, 20 February 2012 (UTC)

Some people believe that nightshade.toolserver.org is down, and that's the reason why the archive bot is not running. Sending a ping to 91.198.174.201 gets 'destination host unreachable'. If the nightshade server is going to be down for a while, maybe someone could move the crontab from there to one of the working toolserver machines. EdJohnston (talk) 03:05, 21 February 2012 (UTC)

I don't know why you say "some people" believe. Nightshade is down and bot operators have been advised afaik to move their cronjobs to willow, which is still operational. As I understand it, the issue has persisted since the extraordinary maintenance on Toolserver on the 16th and 17th, but I might be mistaken on this. Snowolf ^{How can I help?} 07:10, 21 February 2012 (UTC)

Misza has responded at User talk:Misza13#Re: Bots down & stuff. He indicates he will run his archive bots manually beginning at 16:00 on 21 Feb. until somebody fixes nightshade. (He thinks moving the cron jobs to willow could create further problems). The nightshade machine has been down since Friday 17 Feb. In the toolserver issue report, DaB has stated: "Looks like I disconnected the disc-controller or the disc by accident. Needs someone in the colo to fix this." EdJohnston (talk) 14:23, 21 February 2012 (UTC)

Unable to enter text at one site

I wanted to leave a note at talk:Inflation rate world.PNG that the map is in error (It shows the US with 0% inflation, whereas the cited source gives the US's inflation rate as 3%; I didn't check any other countries), but when I twice tried to leave a comment, my heading got erased and I could not (physically) enter any text at all into the text box. Kdammers (talk) 06:02, 21 February 2012 (UTC)

That image is hosted at Wikimedia Commons, which means you will need to use the discussion page there. Here is a direct link where you can enter your comment. --Andrew (User:90) (talk) 06:24, 21 February 2012 (UTC)

Thanks. It's too bad that's not explained (clearly?) at the talk page itself. Kdammers (talk) 07:18, 21 February 2012 (UTC)

It could be explained by a check in MediaWiki:Newarticletext like it already does for the file page. Compare the "Create" tab at File:Inflation rate world.PNG and File talk:Inflation rate world.PNG. You can post a suggestion at MediaWiki talk:Newarticletext. PrimeHunter (talk) 12:55, 21 February 2012 (UTC)

Wikipedia search (2012-02-21)

A query made up of two or more words produced no results until 08:45 of the 21st of February 2012 (UTC). It now runs fine in the main namespace, but not in other namespaces, where it produces no results. Happy editing! –pjoef (talk • contribs) 11:23, 21 February 2012 (UTC)

Same here, manually searching talk archives produces no results. Searching from an archive box produces a lot of results, but clicking "search" again without changing anything produces again no results.

It looks like "/wiki/Special:Search?" works but "/w/index.php?title=Special%3ASearch&" doesn't. Simply replacing this part of the URL fixes the search. I think it that the "?" and the "&" in the query strings and not placed in the correct places when used the second form.

(by the way, the link in the archive box form uses "&fulltext=Search" twice). --Enric Naval (talk) 12:18, 22 February 2012 (UTC)

Changes to Template:Infobox

I don't know if this is the right place, but nothing else came to mind (tell me if I'm wrong), so here goes...

Per this, I'd like to find someone template-savvy who would be willing to sandbox this idea and put the Portal links inside the infoboxes. Nolelover ^{Talk·Contribs} 15:41, 21 February 2012 (UTC)

I hope you plan on getting a wider consensus. I don't think the few people watching WP:PORTAL should be changing something as far-reaching as infoboxes. — Bility (talk) 15:58, 21 February 2012 (UTC)

Well, it was an RfC, but yes, I understand. Hence I'm asking for someone to sandbox this, just so we have some idea what it will look like. FWIW, I couldn't care less where the Portal links go, but since some people thought it was a good idea I figure there's no harm in testing this. Nolelover ^{Talk·Contribs} 16:04, 21 February 2012 (UTC)

For infoboxes using {{infobox}}, simply add the portal to |below=; after gaining consensus. ---— Gadget850 (Ed) ^talk 16:38, 21 February 2012 (UTC)

Reordering account links at the top of all pages

Moved from WP:HD#Reordering account links at the top of all pages

Is there a script that allows me to reorder the links for my account? In particular, I would like to change the ordering from

Toshio Yamaguchi My talk My preferences My watchlist My contributions Log out My sandbox

to

Toshio Yamaguchi My talk My preferences My watchlist My contributions My sandbox Log out

Is that possible for my own account only? Toshio Yamaguchi (talk) 18:15, 21 February 2012 (UTC)

Your sandbox is last? Mine is third (Redrose64 My talk My sandbox My preferences), so I guess the position isn't set in stone. I have no special code for setting it - it's just the default position when I have the gadget 'Add a "my sandbox" link to the personal toolbar area.' turned on. --Redrose64 (talk) 18:22, 21 February 2012 (UTC)

What skin are you using? Grandiose (me, talk, contribs) 18:26, 21 February 2012 (UTC)

Add two more parameters to addPortletLink in your common.js so the last part looks like this: 'Go to my sandbox', null, '#pt-logout'); ; also see the gadget code MediaWiki:Gadget-mySandbox.js. — AlexSm 18:28, 21 February 2012 (UTC)

(edit conflict × 2) Try this:

$( document ).ready( function() {
  $( 'a', '#pt-mytalk' ).text( 'My talk' );
  $( 'a', '#pt-preferences' ).text( 'My preferences' );
  $( 'a', '#pt-watchlist' ).text( 'My watchlist' );
  $( 'a', '#pt-mycontris' ).text( 'My contributions' );
  $( 'a', '#pt-mysandbox' ).text( 'My sandbox' );
  $( 'a', '#pt-logout' ).text( 'Logout' );
});

Hope this helps. (I'm not a great JS programmer! ;) Also note that pt-mycontris is not a typo.) Reaper Eternal (talk) 18:30, 21 February 2012 (UTC)

(edit conflict × 3) He's using Safari, I think it might be a browser incompatibility with whatever method they're using to insert the list node. — Bility (talk) 18:31, 21 February 2012 (UTC)

Alex' solution works. Now 'Log out' is at the far right side and 'My Sandbox' is directly to the left of it. Thanks very much to all for the help. Toshio Yamaguchi (talk) 18:34, 21 February 2012 (UTC)

Clicking on External Link sets off antivirus alert

I was editing Saint Thyrsus and one of the External Links was to www.saintpatrickdc.org etc. When I clicked on the link it set off my antivirus program (ESET NOD32). This has never happened to me before in years of editing. I deleted 2 instances of this link (I was making the references look better), so it's not on the current article page—you'll have to look at older versions of the article. This is the message from my antivirus program: ESET NOD32 Antivirus - Alert Access denied !

Details:

  Web page:
  http://channel-reward-central.com/?sov=146639&id=aDS-cALL-gsociety-dU2FsdGVkX19hZGw3N0lha39LWvtOGQAqbDuDRYN8yrRyEF6lfa9JNQ

  Description:
  Access to the web page was blocked by ESET NOD32 Antivirus.
  The web page is on the list of websites with potentially dangerous content.

www.eset.com

--Kenatipo ^speak! 22:25, 21 February 2012 (UTC)

The domain seems to have been squatted. Reaper Eternal (talk) 22:26, 21 February 2012 (UTC)

Thanks, Reaper. Does Wikipedia do anything about squatted domains? --Kenatipo ^speak! 22:32, 21 February 2012 (UTC)

There's really nothing to be done except remove the offending links per WP:ELNO. Regards, Orange Suede Sofa (talk) 22:39, 21 February 2012 (UTC)

I found a way of fixing the problem. Instead of a URL that ends, for example, with .htm#thyr, if I change it to .shtml, it appears to take me where I want to go. (I can drop the #thyr because the original didn't take me to that section of the page anyway.) An "everything" Search for www.saintpatrickdc.org yields 463 results, almost all of them Catholic saints. It looks like I've been given a project for Lent. Thanks for your help! --Kenatipo ^speak! 00:45, 22 February 2012 (UTC)

I've opened several of the old .htm URLs now and the worst that happens is a "404 page not found" message. My antivirus program may have erred on the side of safety initially, but that's OK with me. --Kenatipo ^speak! 21:02, 22 February 2012 (UTC)

How to add a parameter to a template?

I designed User:Piotrus/TDYK for notifying people about DYKs. I'd like for it to be more specific; I'd like to be able to use a {{subst:User:Piotrus/TDYK|parameter - article's name}}, so that the template would actually mention the name of the article that I specify in the parameter. How can this be done? Perhaps somebody who understands the templates better than me could edit my template and add that functionality to it? In similar vein, I'd like my User:Piotrus/w welcome template to be responsive to a parameter that would specify the name of a WikiProject that the person I am inviting may be interested in joining. --_{Piotr Konieczny aka Prokonsul Piotrus| talk to me} 23:22, 21 February 2012 (UTC)

For example, suppose Template:Testing contains The article name is [[{{{name}}}]]. {{testing|name=Google}} will appear as "The article name is Google". Goodvac (talk) 23:26, 21 February 2012 (UTC)

I've modified your template to include this functionality. Goodvac (talk) 23:31, 21 February 2012 (UTC)

Thanks, I can work with that. Can the name= be omitted, so the template would require just {{testing|Google}} ? --_{Piotr Konieczny aka Prokonsul Piotrus| talk to me} 23:34, 21 February 2012 (UTC)

Yes, you can change {{{name}}} to {{{1}}}. Goodvac (talk) 23:36, 21 February 2012 (UTC)

Javascript for collapse

Where is all the JavaScript for the collapsible boxes on Commons? Specifically, with JavaScript I would like put a hook that executes when the user clicks the box for dropdown. Magog the Ogre (talk) 23:46, 21 February 2012 (UTC)

It looks like commons:MediaWiki:Common.js doesn't have the code for WP:NAVFRAME and implements collapsible tables a bit differently so I recommend you to use this: mw:ResourceLoader/Default modules#jQuery.makeCollapsible. — AlexSm 00:01, 22 February 2012 (UTC)

Technical difficulties

I got the "Sorry! This site is experiencing technical difficulties" notice twice a minute ago, but my edits still went through. Was this just a hiccup, or are there things going on? Ten Pound Hammer • ^{(What did I screw up now?)} 01:12, 22 February 2012 (UTC)

There are constantly things going on, and some may cause an occasional hiccup. Things-going-on log. — Edokter (talk) — 01:25, 22 February 2012 (UTC)

Problem displaying wikipedia content in iPad application

I'm seeing Wikipedia mobile content not sized correctly when displayed within iPad apps using UIWebView. It is as though the mobile version has the page width fixed to the size of the iPad device, not the actual view it is displayed in. This is what I'm seeing: [2] This is what I want to see: [3] This used to work correctly. I suspect Wikipedia has become smarter in detecting it is displaying on an iPad and adjusting the frame size assuming it is being displayed with the iPad Safari browser. Is there any way to specify the content width along the lines of http://en.m.wikipedia.org/w/index.php?title=Jupiter&device-width=320.0 (which doesn't work by the way)? — Preceding unsigned comment added by 24.87.55.100 (talk) 02:03, 22 February 2012

Changing text layout for shorter lines

I'm looking for a solution for someone which will switch page layout from a 'fluid grid style' design to one which adds a max-width statement, to improve legibility for those with difficulty reading long lines. Is there such a tweak, either in .js or .css that I can recommend? Ocaasi ^{t | c} 05:04, 22 February 2012 (UTC)

Add

#bodyContent { width:80% }

to your skin.css and tweak the percentage as necessary. Note that this shortens the width of all pages, including watchlist and contribs pages. You can disable this effect on special pages by prepending .ns-special to the css above (.ns-special #bodyContent...) adding another line of code: .ns-special #bodyContent { width:100% }. Goodvac (talk) 05:25, 22 February 2012 (UTC)

Thanks! Ocaasi ^{t | c} 13:01, 22 February 2012 (UTC)

Oops, I made a mistake with the special pages disabling. Modified post. Goodvac (talk) 18:41, 22 February 2012 (UTC)

Help:My sandbox

I started Help:My sandbox. ---— Gadget850 (Ed) ^talk 12:19, 22 February 2012 (UTC)

Search has gone very bad for several hours.

hopiakuta Please do sign your communiqué .~~Thank You, DonFphrnqTaub Persina. 16:30, 22 February 2012 (UTC)

Special:NewPages queue

Uhh, any idea how the Special:NewPages queue lost about 11 days worth of articles in the course of about 2 hours? See http://toolserver.org/~snottywong/cgi-bin/patrolgraph.cgi?hours=168 —SW— ^yak 01:04, 23 February 2012 (UTC)

According to the time frame you have there, the drop occurs about 3 or 4 hours after the Signpost's special report about page patrolling. We were looking at this today, actually, and have concluded that it was a series of inactive patrollers "waking up" and clearing the queue, thought that may be an incorrect assumption.--Jorm (WMF) (talk) 01:19, 23 February 2012 (UTC)

I suppose the patrol log would reveal the answer, if one was motivated enough to look through it. 11 days of articles is a lot of patrolling for a few patrollers to do though. Kinda makes you wonder if someone just opened up 1000 articles and hit the "patrolled" link without even looking at them. —SW— ^express 16:30, 23 February 2012 (UTC)

I did a little digging and it's not adding up. According to an SQL query on toolserver, only 81 articles were marked as patrolled on February 21st between 19:00 and 20:00 UTC (that's 81 articles in article space only, which were manually marked as patrolled). Yet, the graph shows that the newpages queue lost 6 days worth of articles in that hour. How is that possible? —SW— ^yak 23:38, 23 February 2012 (UTC)

Was this when nightshade was down? See #MiszaBot is down above. --Redrose64 (talk) 12:57, 24 February 2012 (UTC)

Nope, my tasks run on willow. And besides, the newpages queue hasn't been in the single-digits since... well, the last time that mediawiki software was updated and the entire queue got wiped out. The only logical explanation I can think of is that there were a bunch of editors who went on a patrolling spree in the middle of the queue, over a few days, and then when the back of the queue caught up to the hole they made in the middle, there was a big discontinuity in the graph (since the tool which creates the graph only looks at how old the last few articles in the queue are). But still, to erase 11+ days worth of articles from the queue is a pretty monumental task, unless you're barely looking at each article, which is not the point of NPP. —SW— ^{spill the beans} 15:20, 24 February 2012 (UTC)

I did my own analysis day-by-day around the signpost article. It looks mostly like a lot of stuff patrolled by Sfan00 IMG. — Andrew Garrett • talk 00:16, 25 February 2012 (UTC)

Transclusion issue

On WP:TFD, Wikipedia:Templates for discussion/Log/2012 February 16 and Wikipedia:Templates for discussion/Log/2012 February 15 are just showing up as plain links instead of transclusions of those day's logs. Is this just due to template overflow? Ten Pound Hammer • ^{(What did I screw up now?)} 21:18, 22 February 2012 (UTC)

Yep, see Wikipedia:Template limits. At the bottom of the source code is:

NewPP limit report
Preprocessor node count: 57617/1000000
Post-expand include size: 2048000/2048000 bytes
Template argument size: 247794/2048000 bytes
Expensive parser function count: 3/500"

Goodvac (talk) 21:22, 22 February 2012 (UTC)

Perhaps the habitual TFDers could be persuaded to slow down a bit? --Redrose64 (talk) 21:26, 22 February 2012 (UTC)

User page

This probably isn't even an issue for VPT, but it seemed like a place I could get a quick answer; I've always used Firefox, where my talk page looks fine with the discussions left-aligned, but in *cough* Internet Explorer the discussions are centred, which looks terrible. Could someone tell me what the difference is that's causing the problem? Thanks. Black Kite (talk) 01:45, 23 February 2012 (UTC)

In IE, table cells do not inherit the text-align property from the table. And by default, <th> tags are center-aligned. Apply the alignment directly to the cells instead of to the table, and it should work. Anomie⚔ 01:57, 23 February 2012 (UTC)

Also, "text-align:centre" should be "text-align:center". — Edokter (talk) — 01:59, 23 February 2012 (UTC)

Also/or: use "|" instead of "!" at the beginning of the last line of table code, to make it a normal left-aligned <td>...</td> element instead of a bold centered <th>...</th>. In any case, having so many unclosed tables and unclosed table cells is taking chances with the wikitext parser! — Richardguk (talk) 02:06, 23 February 2012 (UTC)

Thanks all. I think "centre" was my UK auto-correct kicking in! Black Kite (talk) 11:34, 23 February 2012 (UTC)

Template:GAN not working Bengali Wikipedia

I have created bn:Template:GAN in Bengali Wikipedia,but it is not substituting in Talk page. Please help us. ---- Jayanta Nath ^{(Talk|Contrb)} 05:00, 23 February 2012 (UTC)

Try replacing "Talk" on the first line with "আলাপ". – Allen4names 06:06, 23 February 2012 (UTC)

Thank you Allen, it is working fine.-- - Jayanta Nath ^{(Talk|Contrb)} 08:14, 23 February 2012 (UTC)

Transclude template without adding category

This may be impossible, but is there a way to transclude a template into a page without that page ending up on a category in that template? (Assuming the template is generally designed to have the page added to a category.) PuppyOnTheRadio talk 12:29, 23 February 2012 (UTC)

You would have to add that feature to the template; see Wikipedia:Category suppression. ---— Gadget850 (Ed) ^talk 12:32, 23 February 2012 (UTC)

Most WikiProject banners have such a feature. If I put {{WikiProject Biography}} on a talk page, that talk page is categorised into Category:WikiProject Biography articles and some others; but if I put {{WikiProject Biography|category=no}}, the categories are suppressed - this is achieved by careful use of {{WPBannerMeta/hooks/cats}} within the project banner.

Outside of the WikiProject banner system, the template {{Category handler}} is used for a similar purpose - see its |nocat= parameter. --Redrose64 (talk) 17:36, 23 February 2012 (UTC)

Was trying to work out a way to do it without altering a template directly. Dagnabbit! PuppyOnTheRadio talk 03:10, 24 February 2012 (UTC)

Mozilla Wiki structure disappears

Firefox 10.0.2: The wiki structure is disrupted. I can't see tabs and the left panel. Images do not have default look. Is anybody else experiencing the same??? Redtigerxyz ^Talk 16:47, 23 February 2012 (UTC)

I can see the left panel in Commons, but the Commons logo on the home page is not visible. Except that everything else looks good in Commons. Redtigerxyz ^Talk 16:50, 23 February 2012 (UTC)

Does this page start with:

The Wikimedia Foundation's 2012 steward election has started. Please vote.

[Hide]

[Help with translations!]

---— Gadget850 (Ed) ^talk 16:54, 23 February 2012 (UTC)

Jump to: navigation, search

The Wikipedia banner (Welcome to Wikipedia)

Today's featured article | In the News

...

• More than 50,000 articles: Bahasa Melayu · Български · Eesti · Ελληνικά · Simple English · Euskara · Galego · עברית · Hrvatski · Lietuvių · ‪Norsk (nynorsk)‬ · Slovenčina · Slovenščina · Srpskohrvatski / Српскохрватски · ไทย

Complete list of Wikipedias

Retrieved from "http://en.wikipedia.org/w/index.php?title=Main_Page&oldid=473360982"

Hidden categories:

Article Feedback Blacklist

Views

• Main Page
• Talk
• View source
• History
• Watch

... I can't see "The Wikimedia Foundation's 2012 steward election has started" banner anywhere on the page in Mozilla, but can see it in IE. Redtigerxyz ^Talk 17:06, 23 February 2012 (UTC)

Sounds like either the stylesheets aren't loading or you have CSS disabled. Do you have Firefox Web Developer installed? Check Firefox View (you may have to press ALT to show the menus) → PAge style. ---— Gadget850 (Ed) ^talk 17:46, 23 February 2012 (UTC)

(undent) Error log from WEb Developer:

• Error: jQuery is not defined

Source File: http://bits.wikimedia.org/en.wikipedia.org/load.php?debug=false&lang=en&modules=skins.vector&only=scripts&skin=vector&* Line: 1

• Error: mw is not defined

Source File: http://bits.wikimedia.org/en.wikipedia.org/load.php?debug=false&lang=en&modules=site&only=scripts&skin=vector&* Line: 1 Redtigerxyz ^Talk 17:16, 24 February 2012 (UTC)

Not getting the problem today. --Redtigerxyz ^Talk 10:28, 25 February 2012 (UTC)

Redoing the code for Wikipedia's fundraiser pages so their contents are readable on Facebook posts

https://donate.wikimedia.org/wiki/Special:FundraiserLandingPage?uselang=en&appeal=Appeal-Sengai

When I post that link on my Facebook page, it should display this preview:

I was born a poor farmer in rural India in 1936. Today I rely on and edit Wikipedia.

I want Wikipedia to be here for all future generations. This is our annual fundraising drive to pay for the servers, small staff and other infrastructure that keeps Wikipedia on the web for free...

But instead, the preview displays the following:

$( document ).ready( function () { // Disable submitting form with return key $( 'form' ).bind( 'keypress', function(e) { var code = ( e.keyCode ? e.keyCode : e.which ); if ( code == 13 ) return false; } ); } ); /* This CSS is responsible for the overall layout of the LP*/ #LP-table {…

What can be done to repair this behavior? Boozerker (talk) 17:18, 23 February 2012 (UTC)

This script tag is inside a paragraph, which could easily confuse Facebook. I've filed this as bug 34660. Superm401 - Talk 23:46, 23 February 2012 (UTC)

Thanks for filing the bug. This is definitely on our radar. We are reworking the entire landing page system for the next fundraiser and hope to fix this and a multitude of other issues in the process. Following bug 34660 is probably the best way to get updates on this. Pgehres (WMF) (talk) 10:40, 25 February 2012 (UTC)

Custom drop down options on special pages

How can I customize personally options found at MediaWiki:Ipbreason-dropdown, MediaWiki:Protect-dropdown, MediaWiki:Deletereason-dropdown and other drop down lists found on various administrative special pages? For example, how could I make "Repeatedly removing speedy deletion templates from own created pages" appear in the Ipbreason-dropdown list for me without actually adding it to Ipbreason-dropdown? I had it suggested that such customization could be done via javascript or something similar, but I'm not sure if this is even possible, let alone how to do it. Any suggestions? Ks0stm ^{(T•C•G•E)} 18:29, 23 February 2012 (UTC)

I tried throwing together some javascript to add additional reasons to the block form (easy enough), but apparently there is some form sanitation code in the MediaWiki engine that blocks any submitted values that aren't on the block reason. Reaper Eternal (talk) 20:03, 23 February 2012 (UTC)

You could make it so when you click a custom reason, it pre-fills the 'mw-input-wpReason-other' textbox then changes the reason to Other. That should be useful and will not be detectable by the server. Superm401 - Talk 00:40, 24 February 2012 (UTC)

I'm a dark background kind-a-guy; I set my browser to ANYTHING but white ... so equations, formulas, practically anything notated at wiki appearing as clear-backgrounded image is absolutely unintelligable

My suggestion is that since those browsers who leave their settings with light background would be unaffected by a change from CLEAR to WHITE backgrounds in these script images, why not make EVERY equation/etc white backgrounded?

That way, despite my dark background, I (capital) would be able to read the darn things. — Preceding unsigned comment added by 67.102.65.21 (talk • contribs) 19:23, 23 February 2012

See Wikipedia:Village pump (technical)/Archive 97#Transparent PNGs used for math formulas unreadable on black background. In short, if you override Wikipedia's styles for the background color, also override img.tex to specify a background color for math images. Anomie⚔ 20:48, 23 February 2012 (UTC)

Alternatively, you can use this script to change the foreground color of all formulae in the current page to any of the supported colors (in the example, it is set to "red"). Helder 21:38, 23 February 2012 (UTC)

Are you using either the background/foreground color selection in the web browser options or the operating system's High Contrast mode to force the black background? Either way, you would only see a white outline around the black text. If I am not mistaken, Wikipedia will, sooner or later, switch to rendering these formulas in JavaScript using MathJax (so there would be no images involved). I know of no good workaround, as most CSS changes do not affect High Contrast mode.

If you happen to use Internet Explorer as your web browser, you could create an account here, log in, and then go to Special:MyPage/common.css and insert the following code to make the formulas white-on-black. Then press Ctrl-F5 to load the new settings.

img.tex { background-color: white; filter: invert; }

Otherwise, just use the script Helder posted, and put it in Special:MyPage/common.js instead (bypassing your web browser's cache afterward). PleaseStand (talk) 04:36, 24 February 2012 (UTC)

Slow wiki?

Is the site really slow for anyone else? It's taking up to 30 seconds for my edits to go through. Ten Pound Hammer • ^{(What did I screw up now?)} 19:33, 23 February 2012 (UTC)

No, this one went through straight away.--Gilderien Talk|Contribs 20:58, 23 February 2012 (UTC)

It was running slow earlier.Edinburgh Wanderer 21:01, 23 February 2012 (UTC)

Article counter

Hi, what happenned with this tool http://toolserver.org/~soxred93/pages/. Exist another similar tool?--Inefable001 (talk) 05:29, 24 February 2012 (UTC)

Here ya go. --Andrew (User:90) (talk) 06:02, 24 February 2012 (UTC)

Basically you can access all of soxred93's tools by replacing the soxred93 part of the URL with tparis. The user that previously operated the tools has retired and TParis now operates them. --Andrew (User:90) (talk) 06:06, 24 February 2012 (UTC)

Ok, thank you--Inefable001 (talk) 07:26, 24 February 2012 (UTC)

permanent link to an older version, found by time code

Hello, is there a way to transfer this scheme:

• …&action=history&year=2007&month=2

for version history to

• &oldidtime= …

for single versions?

The intention is to link a version of the respective page at a currently interesting point in time (e.g. last page version in february 2007, maybe inclusive more accurate parameters) or, similar to FullURL and #expr, to link a version of the respective page [xy] time ago. --Hæggis (talk) 17:42, 24 February 2012 (UTC)

I don't think so. I can't find anything related to dates at mw:Manual:Parameters to index.php, other than filetimestamp, which only works in File: namespace. --Redrose64 (talk) 18:10, 24 February 2012 (UTC)

You can link to a page history only containing the last revision before a given time. See Wikipedia:Complete diff and link guide#Timestamp limits and use limit=1. You cannot link directly to the revision as far as I know. A reader must first click your link and then the single revision in the page history. PrimeHunter (talk) 18:14, 24 February 2012 (UTC)

Red links not go to edit mode

I'm sure this has been asked before, but how do I keep red links from going directly to edit mode? This is more than useless to me and incredibly annoying. —danhash (talk) 20:16, 24 February 2012 (UTC)

One option is to get yourself blocked. Or you could try this userscript to "fix" the red links:

//red links not go to edit mode
$( function(){
  $( 'a.new' ).each( function(i, aa){
    aa.href = aa.href.replace(/[?&]action=edit/, '')
  })
})

— AlexSm 20:37, 24 February 2012 (UTC)

Later fixed to make it work. — AlexSm 22:29, 24 February 2012 (UTC)

Trying for a block isn't necessary. Log out, and a redlink will then go to a generic information page, something like this. --Redrose64 (talk) 21:10, 24 February 2012 (UTC)

I am usually logged-in while using Wikipedia (and for good reason—I edit), so logging out does not fix my problem. —danhash (talk) 21:25, 24 February 2012 (UTC)

Quite: but nor would being blocked, which is the suggestion that I was responding to. --Redrose64 (talk) 21:29, 24 February 2012 (UTC)

Thanks Alex, I'll try it out. I really don't need more JavaScript than is necessary executing on every page load though; I already have enough user scripts that could/should be MediaWiki functions as it is. If I were a casual user it wouldn't be that big of a deal, but especially when I'm working on long pages from moderately-powered computers it gets pretty ridiculous. If MediaWiki can change it for blocked users there should be an option in preferences for everyone. Is a user script the only current way? If so, I propose an option be added by the devs; should be pretty simple. —danhash (talk) 21:04, 24 February 2012 (UTC)

The code doesn't seem to work. I added semicolons and it still didn't work. Is there a typo anywhere? —danhash (talk) 21:59, 24 February 2012 (UTC)

Try this (slightly modified Alex's version):

$( function() {
    mw.util.$content.find( 'a.new' )
   .each( function(i, aa){
        aa.href = aa.href.replace(/[?&]action=edit/, '')
    })
})

Goodvac (talk) 22:15, 24 February 2012 (UTC)

It works! What does the "&redlink=1" part of the URL mean though? —danhash (talk) 22:17, 24 February 2012 (UTC)

It just means that url was a redlink. If you want, I can tweak the code to get rid of that part too. Goodvac (talk) 22:20, 24 February 2012 (UTC)

I figured as much; I guess what I mean is what is the point i.e. what function does it serve? —danhash (talk) 22:22, 24 February 2012 (UTC)

&redlink=1 tells MediaWiki not to automatically open edit mode if you do not have permissions. I'm not sure if there are any other uses. — AlexSm 22:29, 24 February 2012 (UTC)

Yes, it's described on mw:Manual:Parameters to index.php#Optional additional data, bottom of that section. --Redrose64 (talk) 23:12, 24 February 2012 (UTC)

why won't image load on page?

Hi,

I started the page Fred R. Moore, based on an image from the Commons. (The image seems to have moved over to wikipedia.)

But the image won't load. What am I doing wrong? Thanks, MathewTownsend (talk) 21:28, 24 February 2012 (UTC)

Switching to the .jpg version seems to work. I'm not sure what's up with the .tif. Chris857 (talk) 21:35, 24 February 2012 (UTC)

It's not moved over to Wikipedia. You can tell that it's still on commons because of the presence of this box below the resolution information. Commons images are accessible on Wikipedia under the same names - the source is transparent. --Redrose64 (talk) 21:38, 24 February 2012 (UTC)

I don't get what you mean. When I originally linked it, it was on the Commons. I found it by looking through a list of images on the Commons that hadn't been categorized yet. But now when I click the link, its on wikipedia. I can't seem to get back to the Commons version. MathewTownsend (talk) 21:50, 24 February 2012 (UTC)

ok, this link is to the Commons. I went there and got it. But now when I click it, it goes to a wikipedia page. What gives?

MathewTownsend (talk) 21:53, 24 February 2012 (UTC)

never mind, now it's loading. Thanks! MathewTownsend (talk) 21:56, 24 February 2012 (UTC)

The link File:FRED R. MOORE - HUMANITARIAN, EDITOR, LEADER, 1857-1943 - NARA - 535620.tif goes to a Wikipedia page which doesn't actually exist. What you see is a replica of the commons page, with the addition of this box and a few other items near the bottom. In that box, if you click the link titled "description page there", you will get to the true image page on Commons. --Redrose64 (talk) 22:04, 24 February 2012 (UTC)

Yeah, but if I just click the link, it goes to a wikipedia page - even though I linked it to the Commons page. Am I supposed to move it over? Other times I've added an image from the Commons, there hasn't been a problem. MathewTownsend (talk) 22:12, 24 February 2012 (UTC)

The scaled-down versions of the tif won't load for me either. But the full size tif is a ridiculous 8 Mbytes. Have you tried converting to a PNG? — RHaworth (talk · contribs) 22:23, 24 February 2012 (UTC)

Well, now it a jpg file on the article page. Thanks to whoever did that. Really, thanks! MathewTownsend (talk) 22:30, 24 February 2012 (UTC)

I thought that I explained... the link goes to something that only looks like a Wikipedia page. It is a "page" that doesn't actually exist. The image is on Commons. Its description page is on Commons, see Help:File page. Those are the proper places. Nothing needs to be moved anywhere. The whole point about having images on commons is that they are available to be shared by all Wikimedia projects, without having to make local copies on all the different Wikipedias, see Wikipedia:Wikimedia Commons. Images on commons are placed in articles, or linked, using exactly the same syntax as for images held on Wikipedia. There is no special technique, nothing that you can do to the image linking code that means "I want this image to come from Wikipedia", or "I want this image to come from Commons". When you make an image link, it looks for an image of that name on Wikipedia and presents that if available. If there isn't one, it looks on commons and if there is one there, returns the image with the same appearance that it would have had it been held on Wikipedia with the exception that the box MediaWiki:Sharedupload-desc-here is contained within the page. --Redrose64 (talk) 23:00, 24 February 2012 (UTC)

User creation

I have a question about User:Masterg and User:Masterg82. The accounts exist but I don't see their creations in the log (Masterg, Mastgerg82) or in the api, although Masterg82 is showing up as having created his account. Could this be from a registration that happened too many years ago? Or is there some other explanation? Just curious. — Bility (talk) 21:40, 24 February 2012 (UTC)

Hmm. My first thought was that they were created on another Wikipedia, and then SUL'd. But this doesn't seem to be the case - see Masterg and Masterg82. --Redrose64 (talk) 22:16, 24 February 2012 (UTC)

On Special:Listusers, if an entry for a user does not have an attached account creation date, the account was created before the [[new user log was created in September 2005, and also had no edits when the "creation date" feature was added to the special page in January 2009. Graham87 08:02, 25 February 2012 (UTC)

Except for Masterg82 and probably others, for some reason. Graham87 08:05, 25 February 2012 (UTC)

Inputbox with namespaces is broken

The page Help:Contents uses an inputbox with "namespaces=Help**,Wikipedia**,Template**,Category". According to mw:Extension:InputBox, this means the first three checkboxes should be ticked by default. An IP has noted here that this isn't working - by default, no checkboxes are ticked and only the main namespace is searched. -- John of Reading (talk) 22:09, 24 February 2012 (UTC)

 Main

 Help

I don't know why but a comparison [4] of produced code for the above inputbox with namespaces=Main**,Help shows that here at en.wikipedia.org a parameter checked="checked" is not made when a namespace has ** after it. PrimeHunter (talk) 22:50, 24 February 2012 (UTC)

Browsing the code review leads to bug 31158http://bugzilla.wikimedia.org/show_bug.cgi?id=31158. This is fixed in MW 1.19, which is deployed on Meta. ---— Gadget850 (Ed) ^talk 09:56, 25 February 2012 (UTC)

Out of Office Reply to Wikipedia e-mails

I receive a steady stream of e-mails (via the 'E-mail this user' link) on matters which should be dealt with on talk pages. I want to give people who do this a prompt, automatic message about it.

I created a special GMail account and set up an Out of Office AutoReply (GMail is schizophrenic and also calls it a Vacation responder) to give a message saying "if it is a Wikipedia matter, use my talk page".

Seemed to be a good idea - one problem: GMail is sending the AutoReply to the bounce address, wiki<at>wikimedia.org and not to the 'From' address of the sender. Any suggestions?

(Ironically, the first e-mail I received on this new accout was from someone who I had blocked with 'cannot edit own user_talk page' - so the auto message would have been pointless.) — RHaworth (talk · contribs) 22:19, 24 February 2012 (UTC)

Problem displaying panoramic images (Android browser)

Hi, in the Android browser (tablet), Wikipedia pages with very wide images, usually panoramas, are scaled so that the whole image fits the width of the screen, resulting in text so small that it is almost illegible. I wondered if there was anything that could be done at the Wikipedia end to stop this happening.* I know that browser compatibility issues are a pain in the neck... 81.159.106.110 (talk) 04:19, 25 February 2012 (UTC) *Or at the browser end, should anyone happen to know!

User last edit

Hi. I'm trying to generate a list of users who are in a particular user group (ie Admins) that also reflects the date/time of their last edit, so as to work out what admin is on the site at any given time. I'm assuming that there is a way to do this using dpl, but that is unfortunately not within the list of my coding skills. Anyone have any options? PuppyOnTheRadio talk 08:32, 25 February 2012 (UTC)