Wikipedia:Village pump (technical)

 Policy Technical Proposals Idea lab WMF Miscellaneous 

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

The technical section of the village pump is used to discuss technical issues about Wikipedia. Bugs and feature requests should be made at the BugZilla.

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

« Archives, 195, 196, 197, 198, 199, 200, 201, 202, 203, 204, 205, 206, 207, 208, 209, 210, 211, 212, 213, 214, 215

testing templates (continued)

for anyone interested in commenting bugzilla:22135. I would have posted this the previous discussion here, but the discussion has already been archived.

Inline template wikitext formatting

Please take a look at Wikipedia:Centralized discussion/Citation discussion#Inline template wikitext formatting and comment, when you have a chance to do so. Thank you. 3:54 am, Today (UTC−5)

hide sidebar

How do you hide the sidebar?Smallman12q (talk) 02:06, 22 February 2010 (UTC)

What sidebar and which skin? — Bility (talk) 16:26, 22 February 2010 (UTC)

The default sidebar on the left hand side in monobook.Smallman12q (talk) 22:03, 22 February 2010 (UTC)

Well, you probably wouldn't want to hide the whole thing, as the tabs at the top of the page (edit, history, etc.) and the personal links (preferences, watchlist, etc.) are tied up in the same element. So you probably want to hide each element individually. In monobook.css you can use:

#p-logo, #p-navigation, #p-search, #p-interaction, #p-tb {display: none;}

to hide the stuff on the left. Just hiding the elements won't collapse the space over there though – it will just be blank. — Bility (talk) 22:29, 22 February 2010 (UTC)

To also collapse the space, the code #content { margin-left:0; } should work (per Wikipedia:Village pump (technical)/Archive 65#expand the main page contents to full screen?). Svick (talk) 23:16, 22 February 2010 (UTC)

Yes, it works great, thanks! Note: I still had to hide "#p-logo" and "#p-search", which were floating on the side. — Bility (talk) 23:22, 22 February 2010 (UTC)

The wikipedia logo, search bar, and "languages" box aren't hidden, how do I hide those?Smallman12q (talk) 00:49, 23 February 2010 (UTC)

Try this — Bility (talk) 01:20, 23 February 2010 (UTC)

#p-logo, #p-search, #p-lang { display: none; }

Also, you might be interested in user:js/bottomSidebar, from WikiProject User scripts. — Bility (talk) 01:24, 23 February 2010 (UTC)

I got it to work=D. So to "remove" the left sidebar you need to add:

#p-logo, #p-navigation, #p-search, #p-interaction,  #p-tb, #p-lang { display: none; }
#content { margin-left:0; }

Thanks!Smallman12q (talk) 01:51, 24 February 2010 (UTC)

You could also add some js which pushes the sidebar open by undoing the css with a thin, tall, button. I'll try it out... ManishEarth^{Talk • Stalk} 04:57, 24 February 2010 (UTC)

Finished with the script, adding cookie support, ask me for the basic script if u want it. ManishEarth^{Talk • Stalk} 07:53, 24 February 2010 (UTC)

Completely finished. It works perfectly in Monobook, but DO NOT TRY it in vector (beta) as it will push its own tab out of the window. To install, add this to your monobook js page :

importScript('User:Manishearth/sidebartoggle.js');

Then purge your cache. You should see a 'siamese twin' tab (one tab with two links in it) saying "Show sidebar", and "(s)".

The "show sidebar" tab, when clicked, shows the sidebar and turns into a "Hide sidebar" tab. Clicking the "(s)" will save the current configuration (in your browser as a cookie). Basically, the "(s)" makes the current state of the sidebar the default one. ManishEarth^{Talk • Stalk} 08:18, 24 February 2010 (UTC)

Thanks! Works like a charm=).Smallman12q (talk) 01:33, 26 February 2010 (UTC)

Sort issue in Safari 4

Didn't know where best to ask this question, but if anyone from VP could take a look at my question at Help talk:Sorting#Sort issue in Safari 4, it would be appreciated. Thanks.—NMajdan•talk 18:29, 22 February 2010 (UTC)

I'm guessing that the ordering is determined on the first vs. the last element in the table. So in normal sort, it starts at the top, sees that the range is a string, and treats the rest as strings. When you do reverse sort, the algorithm starts at the bottom of the table, sees a number and parser the rest of the tablecolumn as numbers. —TheDJ (talk • contribs) 19:36, 22 February 2010 (UTC)

I have another sort issue, so I'll ask here rather than starting a new thread. A reviewer for an FLC I have up is saying the the postseason columns at List of Texas A&M Aggies head football coaches are not sorting correctly. I've looking in Safari on Windows, and it appears fine. Any help would be appreciated.—NMajdan•talk 16:52, 26 February 2010 (UTC)

Custom edit button: add text, set caret position & auto-summary

Hi, I'm trying to create a custom edit toolbar button that adds a summary when you press it along with adding text to the textarea. I tried various things with javascript but I don't know it very well. Basically, I want it to insert a tag (like {{cleanup}}), set the caret position after the "p", and add the summary "adding {{cleanup}}". The only way I know how to set the caret position is by using the mwCustomEditButtons thing where you can define tag open, tag close, and sample text but I don't think that will work if I want a edit summary to. I'll settle for a bookmarklet if I'm asking too much. I'm using firefox, so I don't need any support for other browsers. Voxii (talk) 00:36, 23 February 2010 (UTC)

Hi, from WikiProject User scripts: you could try User:MarkS/Extra edit buttons or User:Flyingidiot/buttons.js (with some modifications). — Bility (talk) 01:25, 23 February 2010 (UTC)

Thanks, I checked out those pages but neither one mentions anything about an auto-summary. The "with some modifications" part is what I need help with. :) Voxii (talk) 02:16, 23 February 2010 (UTC)

Testing… Svick (talk) 10:25, 23 February 2010 (UTC)

The preceding line was brought to you by

javascript:document.getElementById('wpTextbox1').innerHTML += '\n:::Testing… ~~~~'; document.getElementById('wpSummary').value += 'Testing javascript';

Hope this helps. Svick (talk) 10:28, 23 February 2010 (UTC)

Thank you! It's those little things in js that always trip me up (and online documentation is so bad). I really appreciate it. Regards, Voxii (talk) 12:28, 23 February 2010 (UTC)

We should really work on some good online documentation, with all variables listed, etc. To learn how to add new tabs, fetch page names, etc, I had to prowl the other userscripts. ManishEarth^{Talk • Stalk} 09:15, 24 February 2010 (UTC)

I've suggested that before...the API and overall Wikipedia interface require a lot better documentation (w/ examples=P).Smallman12q (talk) 01:34, 26 February 2010 (UTC)

In a month, when I'm free, I'll make a list of all functions/variables, important stuff, and stuff you should know...ManishEarth^{Talk • Stalk} 03:10, 2 March 2010 (UTC)

Note that the new toolbar and editing interface that the Usability Initiative is working on works completely different from the current toolbar. Its also about a million times more complex and almost completely undocumented. Mr.Z-man 03:57, 2 March 2010 (UTC)

I know about that... I'm first gonna document all variables (eg wgServer, wgCanonicalNamespace) and functions (like addPortletLink()). Then I will work on the edit box for monobook. Beta will come much later. ManishEarth^{Talk • Stalk} 04:29, 2 March 2010 (UTC)

Photo on talk page covers content

I found a solution to this once before but a search of the Village Pump archives didn't show what I had done, at least not on the early pages of the search.

User talk:Scarce has a bear in the last topic. I created a temporary solution, which was the second-to-last topic, but I don't know how to fix it permanently. Also, the topic I created shows that when I added to the topic, the computer I was on didn't show the bear.Vchimpanzee · talk · contributions · 19:25, 24 February 2010 (UTC)

Sorry, what? OrangeDog (τ • ε) 13:56, 25 February 2010 (UTC)

I can't see the bear today. I'm on a computer with Windows 7. The details of my home computer, where I can see the bear, are here.Vchimpanzee · talk · contributions · 21:58, 25 February 2010 (UTC)

What bear? That discussion you linked to is about previous problems with your computer and Yahoo, not with Wikipedia, and nothing about a bear. OrangeDog (τ • ε) 23:05, 25 February 2010 (UTC)

Perhaps it's a bearelephant in the room? --Redrose64 (talk) 23:13, 25 February 2010 (UTC)

Orangedog, you have to scroll down within the discussion to see what my computer is.Vchimpanzee · talk · contributions · 21:48, 26 February 2010 (UTC)

I just checked the page and there's no bear, though there's nothing in the history to indicate it was deleted.Vchimpanzee · talk · contributions · 21:51, 26 February 2010 (UTC)

WHAT BEAR!? OrangeDog (τ • ε) 00:11, 27 February 2010 (UTC)

Was it perhaps this bear or maybe ? Peachey88 ^{(Talk Page · Contribs)} 00:20, 27 February 2010 (UTC)

I think it was the second one.Vchimpanzee · talk · contributions · 19:13, 27 February 2010 (UTC)

Facepalm. OrangeDog (τ • ε) 22:04, 28 February 2010 (UTC)

How to find non-displayed text in WP article?

Hi!

I noticed a (probably) hijacked external link in The Tale of Genji. I'd like to:

1. restore the original link
2. id the vandal

How do I find the text inside a link, i.e. the "spamsite.com" part, in the page history? I've looked at the last several months of page history diffs without finding it. Is there maybe a hidden option for wikiblame?

(I've already commented-out the EL and requested add to blacklist.)

Thanks Saintrain (talk) 21:22, 24 February 2010 (UTC)

The link was added at 1:43, 18 April 2005 by User:Chinasaur, and it hasn't been changed since. The way you find the text (or at least the way I did it) is by editing a previous version. Clarityfiend (talk) 21:41, 24 February 2010 (UTC)

Thanks Clarityfiend. Would you please elaborate on your method? Saintrain (talk) 22:17, 24 February 2010 (UTC)

A simple, but lengthy method is to verify where the link appears in the current version, then check the page history far enough back that you're reasonably certain the link won't be there. For instance, you can check a year back; if the link is still there, check another year back - in this case, you would have to go back a ways. Once you have located a version without the link, and a version with the link, about a year apart, you split the difference. Check a version about six months into the year, and it will either have the link or not, and you have a closer window for when it was added, inbetween your six-month version and one of the two. Then keep on splitting the difference that way until you find it. This requires checking a lot of versions, but it's not really that hard to do. — Gavia immer (talk) 22:38, 24 February 2010 (UTC)

You can automate that using WikiBlame which will compare the versions for you, searching for a specific string of text and when it was added. Regards SoWhy 22:46, 24 February 2010 (UTC)

The sad thing is, I know about that tool, and I still have been doing things manually. Le sigh. — Gavia immer (talk) 22:48, 24 February 2010 (UTC)

I was afraid you might say that :-(. And wikiblame (apparently) locates only "printable" text, so not the link/formatting stuff. Thanks. Saintrain (talk) 00:38, 25 February 2010 (UTC)

It's not so bad, actually - if you've got things down to a year's range, then three careful slices can get you down to a month and a half, six can get you a range of less than a week, and nine get you to less than a day. — Gavia immer (talk) 00:48, 25 February 2010 (UTC)

I've mentioned this several times already to Flominator, he still wont add a checkbox but you can force it by using &tags_present=true [1]. It's almost enough to make me rewrite on the Toolserver. — Dispenser 03:24, 25 February 2010 (UTC)

Sorry, I've added the parameter but forgot to add the checkbox. I will add it soon. --Flominator (talk) 06:51, 26 February 2010 (UTC)

BTW Gavia immer got it exactly right how I tracked the info down. Clarityfiend (talk) 09:38, 25 February 2010 (UTC)

The old binary search gag! But isn't that what computators are for? :-) And thanks Flominator for the cool tool and Dispenser for the (undocumented) &tags_present=true. Saintrain (talk) 18:00, 25 February 2010 (UTC)

You're welcome. Function has been added to the UI now. --Flominator (talk) 18:41, 27 February 2010 (UTC)

After a spam-site is blacklisted is the db then "swept" for the offending site?

And, if so, then what?

Thanks. Saintrain (talk) 00:38, 25 February 2010 (UTC)

Yes, otherwise it would be impossible for anyone to edit articles that link to that site. I'm not sure exactly what you're asking, but Special:Linksearch can be used to find which pages link to a particular website. Graham87 03:25, 25 February 2010 (UTC)

Bug 1505 was fixed back in 2008, so that's not a problem. I don't think there is any automated scanning for the blacklisted links, AFAIK it's up to the people running the blacklist to follow up on the existing links using Special:LinkSearch. Anomie⚔ 04:03, 25 February 2010 (UTC)

That would be correct (or at least the person noticing the spam). MER-C 08:18, 25 February 2010 (UTC)

Thanks for the Special:LinkSearch. Only 79 hits for that domain. That won't be sooooo baaaaaaad. Saintrain (talk) 18:03, 25 February 2010 (UTC)

It doesn't always happen; the Help Desk recently had several requests for help to deal with "unrevertable vandalism" because apparently some websites were put on the blacklist after the vandalism was made but before the wannabe reverter came along. I had to rollback an admin here as if were vandalism, because rollback was the only way to restore a link that had been put on the blacklist even though talk page consensus supported its inclusion on that page. Nyttend (talk) 18:33, 25 February 2010 (UTC)

If rollback bypasses the spam blacklist, that's a bug. If a vandal removes the link, then of course you can't re-add it: that's the point. If you want to re-add it to any page, you have to get it de-blacklisted (as far as I can tell). —Simetrical (talk • contribs) 19:53, 25 February 2010 (UTC)

Add help links to Special:Preferences

Moved from WP:VPR

I have recently been through Meta's meta:Help:Preferences page to bring it into line with today's preferences pages (some stuff was original 2003 content). Each tab in the Preferences page now has a corresponding section heading, and thus a hash link.

I am suggesting that we add a "help" link at the top of each tab in Special:Preferences, maybe worded "Help on/for these options/this tab".

The advantages of this would be:

• Easier for people to get help on what can be a complicated preferences system
• Less need for verbosity on the actual preferences pages themselves
• Reduce helpdesk requests on preferences (currently there is no help link at all on Special:Preferences)

There is a minor issue: the content is on Meta, not here. We have no control over the content, WP-specific info cannot be inserted, and wiki users could get confused about where they have gone if they were sent to Meta. We may need to copy the content locally, or restart the old bot update service that used to run about five years ago, in some form or other.

Any thoughts? — This, that, and the other (talk) 06:25, 25 February 2010 (UTC)

Why would we need to insert enwiki-specific content? There are no enwiki-specific preferences. Also, this material should be at MediaWiki.org, not meta. Fantastic job on getting it up to date, however. Happy‑melon 18:24, 25 February 2010 (UTC)

Strongly agree. These help pages are hard to find, even for experienced editors. There are more people requesting the same thing at Help talk:Preferences#Add link from Special:Preferences, and Wikipedia:Village pump (technical)/Archive 70#Context-related help and searching in a new window.

The Technical issues are, how do we add a link, and where do we add a link, to the Preferences page(s) themselves? I don't know how. As for where, preferably at the top just above the "My preferences" header; or if that is impossible, then at the bottom next to the "Restore all default settings" link. -- Quiddity (talk) 19:48, 25 February 2010 (UTC)

Bogotá Image

The user Jumamuba insist to upload a copyright file. Was deleted on Commons, but many times he put this on the city article. --190.29.129.138 (talk) 12:54, 25 February 2010 (UTC)

Another version. --190.29.129.138 (talk) 12:55, 25 February 2010 (UTC) The deletion request on Commons here.

Yet another copy: File:Bogotacoolpictureb.jpg, both here and on Commons. If you haven't noticed that the file is actually a .bmp, see my new post #Images uploaded with wrong file extension below. PleaseStand (talk) 03:02, 26 February 2010 (UTC)

And another, on Wikipedia (check File links section on description page): File:Bogota skyline and Panoramics.jpg. PleaseStand (talk) 03:20, 26 February 2010 (UTC)

See Wikipedia:Files_for_deletion/2010_January_8#File:BogotaSkyline and another interest points.jpg for what is likely one of the first copies, even though I cannot verify this (deleted file). The user has been reposting pictures for quite a while. Is there any way to stop it? PleaseStand (talk) 03:47, 26 February 2010 (UTC)

Police officer template

How can I modify the Template:Infobox police officer for Charles L. Beck to indicate that he is the 55th Chief of Police for the LAPD -- I think it should be included under his name just like Obama has an indication that he is the 44th POTUS. DRosenbach ^{(Talk | Contribs)} 14:15, 25 February 2010 (UTC)

I'd just edit it in: say: "55th Chief of Police - 11/2009". the template kind of sucks, IMO. However, we could revise it with a 'current rank' section just under the header if you like. — Preceding unsigned comment added by Ludwigs2 (talk • contribs)

You could add something like:

|-
{{#if:{{{chief|}}}|
{{!}} colspan="2" {{WPMILHIST Infobox style|header_bar}} {{!}} {{ordinal|{{{chief|}}}}} [[Chief of Police]]

— Bility (talk) 17:17, 25 February 2010 (UTC)

I'm actually toying with the idea of rewriting the entire template using {{infobox}}, so if we're going to do revisions we might as well commit ourselves to doing proper revisions. --Ludwigs2 18:32, 25 February 2010 (UTC)

Mmmm, makin' templates… — Bility (talk) 19:19, 25 February 2010 (UTC)

alright, I changed it. still needs to prettification, but I'll get to that later if no one else does. --Ludwigs2 21:20, 25 February 2010 (UTC)

Looks good – well done! :) — Bility (talk) 21:32, 25 February 2010 (UTC)

Why is this article sorting wrong?

The article Royal Hospital School sorts under "m". Can someone take a look a figure out why? Thanks. Jason Quinn (talk) 19:38, 25 February 2010 (UTC)

Could you please be more specific? In which categories does it sort incorrectly? Not in those I checked. Hans Adler 19:45, 25 February 2010 (UTC)

Fixed. — Bility (talk) 19:57, 25 February 2010 (UTC)

Thanks! Geez, I even looked for that and missed it. Good eyes! Jason Quinn (talk) 20:01, 25 February 2010 (UTC)

User Registration

Seems to be broken at the moment, admins have been notified. Q ^{T C} 21:40, 25 February 2010 (UTC)

Resolved

Seems fixed now, was down for about 1:45. Q ^{T C} 21:51, 25 February 2010 (UTC)

Bouncing edit window

Perhaps this should be on the computing reference desk since it was also happening in an email. But I find that the text I am editing bounces. In the email, when I would click on one of the other activities I was working on, the email's editing box would jump so I could only see the top of the message. It's not as serious on Wikipedia, but it is annoying. I have to keep scrolling back down to see what I'm working on, though at least what I'm typing stays on screen, which is more than I can say for one message board site where the problem is separate but similar and has been happening a long time.

I'm on a computer with Windows 7.Vchimpanzee · talk · contributions · 22:01, 25 February 2010 (UTC)

More: The gray rectangle on the right side, whose size is supposed to vary with the fraction of the entire page or section being edited, goes from tiny to huge (almost as big as the entire right side of what is visible) and back again as I type the edit summary.Vchimpanzee · talk · contributions · 22:08, 25 February 2010 (UTC)

Perhaps screenshots would work better. I can't understand from your description. —TheDJ (talk • contribs) 02:19, 26 February 2010 (UTC)

It would have to be a video. I'm not sure, even if I could remember how to do screenshots, that it would show what is happening. And I won't be on that computer for another week. That day I won't have much time.Vchimpanzee · talk · contributions · 21:47, 26 February 2010 (UTC)

Odd search bug

Something in this article is causing an odd bug in my browser when I try to do a search on the article William Halsey, Jr.. Using IE7.0.5370.13 with Windows XP MCE (SP3), I get "An error has occured in this dialog. Error: 239 Object doesn't support this property or method" This is the only article that does this, I've tried closing the tab and the window, but the behavior sticks to only this article. bahamut0013^words_deeds 23:38, 25 February 2010 (UTC)

Odd indirect transclusion of db-g6

Currently, several talk pages show up in CAT:CSD for no obvious reason. For example Talk:University of Medicine and Dentistry of New Jersey. If you edit this, among the list of "Pages transcluded onto the current version of this page:" are:

• Template:Db-g6 (edit) (semi-protected)
• Template:Db-meta (edit) (protected)

But as neither is directly transclsuded, they must be indirectly transcluded via one of the many other templates on that page. Is there any way to find out which one short of manually examining the code for every template called, directly or indirectly, from this page? DES ^(talk) 00:25, 26 February 2010 (UTC)

Other examples:

• Talk:William Cowper Alexander
• Talk:Wall Township, New Jersey
• Talk:Warren Township, New Jersey
• Wikipedia:WikiProject New Jersey/Deprecation of comments

There seems to be a NJ connection. DES ^(talk) 00:34, 26 February 2010 (UTC)

(edit conflict) Seems to be related to Wikipedia:WikiProject New Jersey/Deprecation of comments. The transcluded comments pages have now all been speedied, so the talk pages should be out of the CSD cat as soon as the job queue gets round to doing it. A null edit will remove a page immediately if that's needed for some reason. Algebraist 00:36, 26 February 2010 (UTC)

As long as I know, and it is going away, no problem. I think i understand now what happened. DES ^(talk) 00:37, 26 February 2010 (UTC)

PHP json decode

Resolved

 – Smallman12q (talk) 11:51, 26 February 2010 (UTC)

I'm trying to print a nested variable ("url") from a json response in php 5.3.

Json code

{
  "offset": "0",
  "results": [
    {
      "body": "body info",
      "date": "date info",
      "title": "title info",
      "url": "url info"
    }
  ],
  "total": 1
}

My current php code

$info = json_decode($jsonresponse);
print $info->{"results"}->{"url"};//this doesn't work

How would I print the value of the "url" variable in php? (This should be fairly simple...)Smallman12q (talk) 02:44, 26 February 2010 (UTC)

It should be

$info = json_decode($xml,true);
print $info["results"][0]["url"];

Smallman12q (talk) 11:51, 26 February 2010 (UTC)

Images uploaded with wrong file extension

It is apparently possible to upload a .bmp with the .jpg file extension (apparently by setting the upload title), and the server generates a .png preview. (See the MIME type of File:Bogotacoolpictureb.jpg, which I have identified as a repost of non-free content and might be deleted soon.) Because we shouldn't use BMP files here since they are a waste of disk space, I am wondering: how many files here are uploaded under the wrong file extension? PleaseStand (talk) 02:54, 26 February 2010 (UTC)

If I remember correctly, you get a warning when you do that. And we can move those files to a correct title + extension. —TheDJ (talk • contribs) 13:29, 26 February 2010 (UTC)

I compiled a list of files with mismatched MIME type and extension and files with extensions we may not want (like bmp) or i don't know (like graffle and xcf). It is available at [2]. Svick (talk) 14:57, 26 February 2010 (UTC)

Oh nice list. I'll plough trough some of that. Do we have any CSD rules about incorrect filetypes ? I don't think so right ? —TheDJ (talk • contribs) 15:32, 26 February 2010 (UTC)

Its probably never come to the notice of anyone. Either request a CSD rule/make one yourself. Or get bmps approved (I'm against it, whoever makes an image should have enough time to save it as jpg). Or request a tweak in the mediawiki code (read here).

I checked out the wp upload settings here. Search for "wgFileExtensions", and you'll see this (see the "read here" link above to understand the variables):

/**
 * This is the list of preferred extensions for uploading files. Uploading files
 * with extensions not in this list will trigger a warning.
 */
$wgFileExtensions = array( 'png', 'gif', 'jpg', 'jpeg' );

/** Files with these extensions will never be allowed as uploads. */
$wgFileBlacklist = array(
    # HTML may contain cookie-stealing JavaScript and web bugs
    'html', 'htm', 'js', 'jsb', 'mhtml', 'mht', 'xhtml', 'xht',
    # PHP scripts may execute arbitrary code on the server
    'php', 'phtml', 'php3', 'php4', 'php5', 'phps',
    # Other types that may be interpreted by some servers
    'shtml', 'jhtml', 'pl', 'py', 'cgi',
    # May contain harmful executables for Windows victims
    'exe', 'scr', 'dll', 'msi', 'vbs', 'bat', 'com', 'pif', 'cmd', 'vxd', 'cpl' );

/** Files with these mime types will never be allowed as uploads
 * if $wgVerifyMimeType is enabled.
 */
$wgMimeTypeBlacklist= array(
    # HTML may contain cookie-stealing JavaScript and web bugs
    'text/html', 'text/javascript', 'text/x-javascript',  'application/x-shellscript',
    # PHP scripts may execute arbitrary code on the server
    'application/x-php', 'text/x-php',
    # Other types that may be interpreted by some servers
    'text/x-python', 'text/x-perl', 'text/x-bash', 'text/x-sh', 'text/x-csh',
    # Client-side hazards on Internet Explorer
    'text/scriptlet', 'application/x-msdownload',
    # Windows metafile, client-side vulnerability on some systems
    'application/x-msmetafile',
    # A ZIP file may be a valid Java archive containing an applet which exploits the
    # same-origin policy to steal cookies
    'application/zip',
);

/** This is a flag to determine whether or not to check file extensions on upload. */
$wgCheckFileExtensions = true;

/**
 * If this is turned off, users may override the warning for files not covered
 * by $wgFileExtensions.
 */
$wgStrictFileExtensions = true;

/** Warn if uploaded files are larger than this (in bytes), or false to disable*/
$wgUploadSizeWarning = false;

It's obvious after reading the manual that bmp isn't blacklisted in $wgMimeTypeBlacklist, so there is no effect. See if you can get it changed. In the meantime, create a wikiproject to download, convert, and reupload those images. ManishEarth^{Talk • Stalk} 16:09, 26 February 2010 (UTC)

I have a wikilist here: User:TheDJ/badfiletypes for those that want to assist. —TheDJ (talk • contribs) 16:37, 26 February 2010 (UTC)

bmp wouldn't be blacklisted. Though we discourage the use of suboptimal file types, we don't block them since encouraging collaboration is generally a greater good than the savings in bandwidth / storage, especially since we can clean up those issues later without too much trouble. Regarding the general phenomenon of incorrect extensions, Mediawiki is planning to move towards an extension-less system so rather than having "File:Bogotacoolpictureb.jpg" we would just have "File:Bogotacoolpictureb" whose image type could change over time and the server would transparency take care of delivering the right extension and mime type information to browsers. The fact that one can change the content type of an existing image now and the server doesn't barf is part of that, but the implementation to fully eliminate extensions is not yet complete. Dragons flight (talk) 16:46, 26 February 2010 (UTC)

From experience, transparent PNG images with a JPG extension do not render properly. ---— Gadget850 (Ed) ^talk 16:52, 26 February 2010 (UTC)

There will always be people who believe that renaming Example.bmp to Example.jpg is enough, and don't appreciate the process of changing the file type when doing a "Save As". It should be possible to verify the file format at upload time; I believe that many common image file formats have a "magic number" - two bytes at the beginning which identify fhe file type, which can be more reliable than the file extension. It's not confined to UNIX - files conforming to the Microsoft BMP format, for example, always begin with the hex data 42 4D ("BM"); JPEG files begin FF D8. --Redrose64 (talk) 17:23, 26 February 2010 (UTC)

There is already a check for the type when moving a file— you cannot move a .PNG to a .JPG (try it). There is no check when using the upload a new version link. This is actually handy when you want to upload a new version of the same image in a different format— all you have to do is then move it to the proper extension. You don't have to orphan the old version, you can keep the old versions if you mess up and you don't need to change the rationale and licensing. —Preceding unsigned comment added by Gadget850 (talk • contribs) 17:53, 26 February 2010

I meant that users could be renaming files on their home computer prior to upload, and not on Wikipedia afterward. They see that message "Change the destination filename to something descriptive. Don't use default filenames" and later on "Permitted file types: png, gif, jpg, jpeg, xcf, pdf, mid, ogg, ogv, svg, djvu, oga."; they may not appreciate what the file extension can be used for, and rename "doc1.bmp" to "Cool image.jpg". --Redrose64 (talk) 18:24, 26 February 2010 (UTC)

The main problem with MWs extensionless version is that there are already multiple images which exist in multiple formats and they are different. (Example: File:Americanflag.JPG and File:AmericanFlag.png.. completely different). In the meantime, MW could make sure that there is no MIME 'phishing'. Actually, BMP's are a small problem compared to redirects, see discussion here. That discussion didn't go anywhere because people were confused on the proposal and how to implement it. I'm working on an airtight method to for this to work, once i get time (in about a month). ManishEarth^{Talk • Stalk} 02:19, 27 February 2010 (UTC)

←I wrote a Toolserver script to display files with incorrect MIME types; it's available at tools:~earwig/cgi-bin/file_extensions.py. — The Earwig ^(talk) 17:17, 27 February 2010 (UTC)

I made a template {{wrongmime}} which is used like this: {{Wrongmime|.abc}} to produce

Template:Wrongmime ManishEarth^{Talk • Stalk} 03:21, 1 March 2010 (UTC)

I've just come across an instance of incorrect extension as I described earlier - File:File-Little Lost Hen.jpg is a Word doc. --Redrose64 (talk) 12:29, 1 March 2010 (UTC)

No need to report it here. User:Earwig wrote a toolserver script here, which finds all filed with incorrect MIMEs. You can also config it to process a certain MIME type. If you are interested in checking these images, fixing them (by downloading, using "save as", and re-uploading), or marking them for deletion, please form a Wikiproject (Be sure to list it at Cleanup wikiprojects directory). Maybe WP:Wikiproject FixImage? ManishEarth^{Talk • Stalk} 13:14, 1 March 2010 (UTC)

preferences and signatures

Would it be possible to design a tool to be self activate in the my preferences section that reminds an editor when he or she has forgotten to sign something? Given that we have a similar tool to remind people about providing edit summaries I would think this would be something easy to tweak, and may help ease the work load of sinebot at the same time. TomStar81 (Talk) 10:13, 26 February 2010 (UTC)

The problem is that signing is very context specific. As such, letting the code interpret this on every post is complicated, and even worse, probably project specific. Such things are difficult to solve in the core software. —TheDJ (talk • contribs) 13:31, 26 February 2010 (UTC)

Ain't that hard, a JavaScript one-liner gadget could look for the four tildes. — Dispenser 18:30, 26 February 2010 (UTC)

I agree, it wouldn't be too hard to implement. If TheDJ is talking about making the the tool "correct"—meaning it only tells you to sign if you need to and deals with all the various circumstances that we encounter when signing or not signing an edit—then I agree; that's much too complex for a computer to try solving on every post. If all TomStar81 wants is something like a not signed or signed under the Save page button or next to the edit summary preview, it would be doable based on the presence of the four tildes in the new text. As long as it doesn't impede editing, I think some kind of notification could be helpful. — Bility (talk) 19:14, 26 February 2010 (UTC)

The problem is not so much one of detecting an unsigned post, it's knowing which unsigned posts should have been signed. A general rule is that all talk page posts should be signed, and article space posts should not. Unfortunately there are some namespaces (such as Wikipedia:) where posts are normally unsigned, but in specific cases (such as Wikipedia:Help desk and Wikipedia:Village pump (technical)) they should be signed; so a simple all-encompassing rule can't be put together. --Redrose64 (talk) 19:34, 26 February 2010 (UTC)

The general rule could be all namespaces, or all namespaces but 0. We're just talking about a simple, unobtrusive notice. Its meaning should be that you haven't signed, not that you should sign. — Bility (talk) 19:36, 26 February 2010 (UTC)

I was using "should" in its imperative form. --Redrose64 (talk) 20:05, 26 February 2010 (UTC)

I agree with you. :D Determining whether a user should sign in every scenario is too complex, which is why I never suggested it. I'm talking about a simple indicator of whether the user has signed. It would then be up to the user's human brain to decide whether they should sign. — Bility (talk) 20:15, 26 February 2010 (UTC)

Like Bility said, we do not need a tool to determine whether we should sign, this would just be for cases like in the talk name space and then only for when folks forget to add four tildes to sign. As a practical matter, the tool does not need to work in every case; what this tool fails to pick up Sinebot should detect and fix accordingly. TomStar81 (Talk) 22:39, 26 February 2010 (UTC)

The other problem is that sometimes you fix a part of the discussion like done here by Bility when I forgot to link the image. The script should prompt first. Why not ask the operator of SineBot for the algorithm? ManishEarth^{Talk • Stalk} 04:05, 27 February 2010 (UTC)

Check User:SineBot#What it looks for. Note that SineBot sees the code after the tilde-based signature has been expanded as well as all substed templates. So therefore it has to look for the two components of a proper signature in any edit to a talk page: the link to the user page and the UTC timestamp, according to that page. Unfortunately, for user JS to implement the same approach, it has to somehow have access to a diff, and to implement it in the same way as SineBot (to account for self-signing templates), that diff would need to come from the server. Here's some code that has been around for several years, but if you both read the (rather messy) script and understand my explanation above, you know why it is limited to only a specific set of pages. PleaseStand (talk) 04:39, 27 February 2010 (UTC)

Reference

Winter_storms_of_2009–2010#cite_note-51 seems to go into the next column in my firefox browser...is it just me or is this a bug?Smallman12q (talk) 14:31, 26 February 2010 (UTC)

That is correct, it is a very long unbreakable string, and thus it overflows the page. —TheDJ (talk • contribs) 14:51, 26 February 2010 (UTC)

mygiganticstringisprettycoolbecauseitmakesthispageevenwiderthanitshouldbeanditstotallyunreadableaswellsoitshouldbefinetoincludeinthispageandwillbeveryusefultoanypotentialfuturereaderswhoreallyshouldntcareaboutseeingthisurlwithaproperdescriptionaswouldbepossiblebyusingthebracketnotation. Basically the url is one long word :D —TheDJ (talk • contribs) 14:55, 26 February 2010 (UTC)

The URL also seems to be a redirect to another target. There are three ways of fixing this:

• replace the exceedingly long URL with the true target, which is hopefully shorter
• place square brackets around the URL and give it a title
• use the {{cite web}} template

The first option is the least preferable because a bare URL is not very intuitive. See Wikipedia:Linking#External links. --Redrose64 (talk) 15:15, 26 February 2010 (UTC)

Especially since the link in question is a Google link that redirects to a blog home page. ---— Gadget850 (Ed) ^talk 17:48, 26 February 2010 (UTC)

Please don't "fix" HTTP redirect, often the new URL has less Wayback history. For the semi-productive editor there's WP:REFLINKS, for the lazy editor the a one-click bookmarklet, that runs a bot. — Dispenser 18:37, 26 February 2010 (UTC)

So add the archive link at the same time as fixing the url. Problem solved. OrangeDog (τ • ε) 13:09, 27 February 2010 (UTC)

Template documentation

Why does the left border of the template documentation of templates such as {{R from alternative name}} and {{Copy to Wiktionary}} not match (on IE8) the left border of the documentation transclusion notice ("The above documentation is transcluded...")? Compare to, for instance, {{Dicdef}} and {{Notability}}. Thanks, -- Black Falcon ^(talk) 21:12, 26 February 2010 (UTC)

IE7 also, but not Firefox 3.0.18 (in IE the upper green box is wider than normal; the lower green box is normal width). --Redrose64 (talk) 21:19, 26 February 2010 (UTC)

Hmm, the upper element is a div, and the bottom element is a table. I'm guessing that is causing the difference... I'll point David at it. —TheDJ (talk • contribs) 21:32, 26 February 2010 (UTC)

Should take this to Template talk:Documentation. ---— Gadget850 (Ed) ^talk 02:48, 27 February 2010 (UTC)

In the interest of keeping the discussion in one place, I've posted notices linking to this discussion at Template talk:Documentation and Wikipedia talk:WikiProject Templates, though I suppose that this thread could be archived and moved if needed. -- Black Falcon ^(talk) 06:10, 28 February 2010 (UTC)

Mobile Wikipedia

At http://en.m.wikipedia.org , you will find a page which is effectively the main page for mobile telephony. It only contains information that changes on a daily basis. The expectation is that mobile traffic will bring the biggest growth to Wikipedia and, as the English Wikipedia grows most in absolute numbers, it is imho important that its main page gives a good mix of information. For me it means that there is at least a link to a page where a lot of the information that can be found on the standard "Main page" can be found..

A "How to" was published on the Wikimedia Techblog and the Swedish Wikipedia is where new functionality was developed. My problem is that I am no good at templates and stuff so I am looking for people who can make something awesome for the English Wikipedia and, this should be the example that other Wikipedias can copy.

So, my request is one that allows you to help yourself (ie en.wp) and, when this is done well, it will be of benefit to many more Wikipedias. Thanks. GerardM (talk) 12:16, 27 February 2010 (UTC)

Another problem with the mobile site is that infoboxes take up too much room. On normal WP, infoboxes are on the side, so you don't notice their length, but on the mobile site, infoboxes are shown at the top, and they require a lot of scrolling. They should be shown by default but have a show/hide button like the sections on the mobile site. Most WP editors don't use the mobile site, so its kinda neglected.. ManishEarth^{Talk • Stalk} 16:10, 27 February 2010 (UTC)

I'd prefer 'slow development' as a description :D —TheDJ (talk • contribs) 18:10, 27 February 2010 (UTC)

The mobile is there because the staff got iPhones/iPod touchs and wanted something better. Same goes for that apple-touch-icon and other vendor specific stuff the enwiki community would object to (see previous discussion on webslices). We can't even edit what goes on the homepage and the stylesheet is hack together of whatever looks good on Safari. The actually mobile/lite interface was apparently suppose to be the chick skin (as far I've been able to gather). — Dispenser 21:14, 27 February 2010 (UTC)

That's not entirely true. It also works on Opera Mini, Sony Ericsson phones, PSP, Nintendo Wii and a few other devices. You are more than welcome to create more refined stylesheets. And you very much can edit what goes on the homepage (if you built a custom portal page). Just let me know when you got something or file a ticket in bugzilla. Also, if you have a good icon for the mobile version, that would be appreciated as well. The chick skin is a slim skin for mobile devices, but it is a tad cluttered and cannot very well adapt to multiple devices. That doesn't mean it won't be used in some form. —TheDJ (talk • contribs) 21:35, 27 February 2010 (UTC)

Viewing a 'cached' page

Hello everyone

I've just noticed this over the last month or so. Whilst viewing Wikipedia my computer now seems to be bringing up cached pages for whatever reason. I've tried disabling viewing cached pages in my preferences but so far no difference.

The reason I think it's a cached page I'm viewing is because on the page it always says "retrieved from".

Does anyone have any ideas regarding this? 5 albert square (talk) 15:02, 27 February 2010 (UTC)

Do Wikipedia pages look normal otherwise? Which browser and skin are you using? It sounds like your browser isn't properly loading/interpreting the Wikipedia style sheets. Try to empty your browser cache and check your CSS/JavaScript settings. Graham87 15:42, 27 February 2010 (UTC)

Hi Graham, thanks for that. it's my works computer that Wikipedia looks strange on so I don't think it's anything to do with skins etc as Wikipedia looks normal on my home pc. I will try to clear my cache and check my Java settings next time I'm with the pc and will report back :) --5 albert square (talk) 13:29, 28 February 2010 (UTC)

Most convenient way to compare sandbox to production?

For templates with a sandbox version, I often want to compare the sandbox version to the production version to see if there are any pre-existing changes changes before I test one. I'm doing it clumsily now: I copy the production source, edit the sandbox, replace all the text by pasting from the clipboard, and then click "Changes". Is there a better way, and in particular, is there a way to get WP to compare the two pages? If so, can one of the template documentation templates be modified to include such a link? — John Cardinal (talk) 15:20, 27 February 2010 (UTC)

When you go to the diff page, in the url (web address) you will see at the end there is a string which looks like "&diff=x&oldid=y". The y value is the page version on the left, and the x the page version on the right. Each edit has it's own "version number". So you can simply copy the most recent version number of the production page, and compare it to the most recent for the sandbox page. Hope that helps. - Kingpin¹³ (talk) 15:26, 27 February 2010 (UTC)

Wow... does that work for any pair of pages? Say there are two articles with a lot of similarity except a few names, could I compare them in this way? --Redrose64 (talk) 15:42, 27 February 2010 (UTC)

Indeed it does ! —TheDJ (talk • contribs) 15:57, 27 February 2010 (UTC)

Thanks. I didn't know that page versions were unique across pages. Makes sense, I guess; I assume the version number is the DB's primary ID for pages.

Is there a way for a template to get the current version number of a page so that it could build the diff link automatically? — John Cardinal (talk) 16:23, 27 February 2010 (UTC)

Left-hand side, toolbox, permanent link (in Monobook, anyway; different skins may place this elsewhere). --Redrose64 (talk) 16:30, 27 February 2010 (UTC)

Wanna script? Here you go:

function getID(){
if(document.getElementById('t-permalink')){
return document.getElementById('t-permalink').innerHTML.split("&oldid=")[1].split("\"")[0]
}else{return "No id" 
} 
}
addOnloadHook(
function(){
var diffPortlet='p-cactions'
/*can be:
p-navigation for navigation sidebar
p-tb for toolbox
p-interaction for interaction sidebar
*/
if(window.location.href.indexOf("action=edit")==-1){
addPortletLink(diffPortlet,"javascript:prompt('The Id for this page is:',getID())",'ID', 'ca-id','Page ID', '')
addPortletLink(diffPortlet,"javascript:document.location='http://en.wikipedia.org/w/index.php?title=Main_Page&diff='+getID()+'&oldid='+prompt('Enter ID for other page','')",'FlexDiff', 'ca-diff','Get diffs for this page with any other page', '')
}
} 
)

Add it to your monobook js page (go to vector.js if you use beta) and purge your cache. You'll get two new tabs (or links in your chosen portlet if you change the diffPortlet variable). In beta, the tabs will be hidden under a little down arrow next to the history tab and 'watch this page' star.
To get a diff, first open the page you want on the left of the diff, and click ID. Copy the ID, and open the page you want on the right. Click the FlexDiff button (It should be right next to/under the ID tab), and it will ask for the ID that you copied. Paste it and click OK. The diff will open in the same window (no need for a new window as the script doesn't work on edit pages). Enjoy!
P.S. Isn't it weird that the title parameter in diffs doesn't matter? this has title=Main_Page, while this has title=Wikipedia:Don't_stuff_beans_up_your_nose. Both link to the same diff.ManishEarth^{Talk • Stalk} 03:26, 28 February 2010 (UTC)

Thanks. I'll experiment with that. What I was thinking, though, was if a template could get the ID number of a page, then the sandbox page or the doc page could get the two IDs of interest and include a link to compare them. Knowing you can compare different pages helps a lot, but I was looking for the icing on the cake. — John Cardinal (talk) 04:28, 28 February 2010 (UTC)

Well, there is a magic word for the revision id. It's {{REVISIONID}}. For example, the revid for this current page is -. But I don't think there's any way to get the revision id of another page without adding a magic word which uses AJAX (I doubt this will happen).ManishEarth^{Talk • Stalk} 08:43, 28 February 2010 (UTC)

Why would it have to use AJAX? {{PAGESIZE}} doesn't. Template:Bugzilla. OrangeDog (τ • ε) 21:55, 1 March 2010 (UTC)

PAGESIZE gets the page size of the current page. Every time you save, the displayed value is updated. Same with REVISIONID. But, to fetch the ID of another page, you either have to use AJAX or update the displayed value on all pages with the magic word every time you save.

Page view statistics function

The article page view statistics function is down for the past three days. Any idea why? Thanks. Dr.K. ^λogos_πraxis 03:06, 28 February 2010 (UTC)

Small request from Bengali wikipedia

This is a small request from Bengali wikipedia. I know that is not a right place for that. We have a page named LocalSettings.php. I am admin there. This page is a main namespace. We know that it is not an article but its occupied one article count. There is no page like LocalSettings.php in this english wiki. If I delete this file, could it create technical issue/problem for this delete? Any one know that what is for that? Main creator cant memorized why he create that page. Please help us. - Jayanta Nath ^{(Talk|Contrb)} 15:39, 28 February 2010 (UTC)

You can delete it without problem. Changes to LocalSettings.php can only be made by the wikimedia system administrators. Such changes should be requested at bugzilla:. The page you name cannot affect the software. —TheDJ (talk • contribs) 15:45, 28 February 2010 (UTC)

Thanks you for this confirmation.- Jayanta Nath ^{(Talk|Contrb)} 15:54, 28 February 2010 (UTC)

Creating a category redirect

I am redirecting Category:Infections of the skin and subcutaneous tissue to Category:Infection-related cutaneous conditions, but I want to do so such that the Category:Infections of the skin and subcutaneous tissue is not itself categorized. Restated, when I look at the Category:Infection-related cutaneous conditions, I do not want want "Infections of the skin and subcutaneous tissue" listed. Any ideas? ---kilbad (talk) 16:58, 28 February 2010 (UTC)

It looks fine to me. However, I note that you created the redirect in the incorrect manner (#REDIRECT [[]] cannot be used for categories); and a 'bot has altered it to the {{Category redirect}} template. Perhaps you last checked it before the 'bot edit? --Redrose64 (talk) 17:19, 28 February 2010 (UTC)

I had... looks great now! Thank you. ---kilbad (talk) 18:19, 28 February 2010 (UTC)

Proposal to add Template:Inappropriate comment class to MediaWiki:Common.css

/* Makes boxes partially transparent on mouse over in [[Template:Inappropriate comment]] */
.InappropriateCommentMouseOver {
    opacity: 0.9;
    filter: alpha(opacity=90);
}

.InappropriateCommentMouseOver:hover {
    opacity: 0.1;
    filter: alpha(opacity=10);
}

Template:Inappropriate comment has a "tag over" feature that places a semi-transparent box over an offensive comment. It's a stern step before refactoring or removing the comment under WP:TPG. When a mouse hovers over the semi-transparent box, this class should decrease the opacity from 90% to 10%. The hover feature has already been discussed on the template's talk page, please note. What do you guys think? ChyranandChloe (talk) 17:14, 28 February 2010 (UTC)

Isn't that very browser specific ? I'd think that many older browsers won't support this right?. —TheDJ (talk • contribs) 17:21, 28 February 2010 (UTC)

My buzzword of choice would be WP:NOTCENSORED; although as always the point can't be summed up with one link. Using this method to hide comments is not saying that you think the comment is inappropriate, it's saying that you don't think other people should (have to) read it. That's not the same thing at all. If a comment is inappropriate, it should be removed; and replaced with a "redacted" note if the removal disrupts the flow of the discussion. The only difference between a disruptive talk page comment and an image of Mohammed is that the presence of disruptive talk page comments hampers our ability to create an encyclopedia; as such, they should be removed entirely, not just censored. There's also a healthy dose of WP:DENY involved; you're creating a huge "hey look guys, I wrote something bad, aren't I bad" flag, rather than achieving the goal of purging discussions of disruptive comments. Is there a discussion which concluded that hiding disruptive comments in this fashion was a good idea? Because I'm thoroughly unconvinced. And that's before we even get near the technical issues TheDJ raises, which are indeed serious (you know the box will be permanently opaque on FF<2, IE<=6, and various other browsers?) Happy‑melon 17:37, 28 February 2010 (UTC)

Two issues then. I use browser shots with this as the test url. Works for Firefox 0.9, IE six and up, and most others. Compared to .reflist columns, which doesn't work for the number one in browser shares, IE, I think it fairs.

Yeah, there's been discussion. As much as I'd like to agree with you, if removing a comment were so easy, then there wouldn't be debate at the climate change probation about it. You should ask 2over0, he's admin who worked on there. ChyranandChloe (talk) 06:57, 1 March 2010 (UTC)

Accessibility looks reasonable. Old browser degradation is acceptable I think. The template would require a namespace check in my opinion, but besides that, I see no huge technical problems (at least no bigger problems than collapse has now). Question becomes, would the community accept it ? —TheDJ (talk • contribs) 14:19, 1 March 2010 (UTC)

(outdent) I don't know. While I was brainstorming features last month, this was the plan: an experiment. I can track which pages use "tag over" either through categories or what links here. After randomly selecting sixty uses (to satisfy normality for a placebo controlled two sample t-test), I can assess whether it's working as a stern warning or failing as a flawed flag. People are affected by their own philosophy. This makes judgement objective.

Of course it works better with all the features in. I appreciate all the attention you've given, but please remember that this discussion isn't about whether "tag over" is acceptable, it's whether a CSS class can be added to MediaWiki:Common.css. The acceptability of "tag over", in my opinion, should be discussed on the template's talk page. Thanks! ChyranandChloe (talk) 05:32, 2 March 2010 (UTC)

Question

I am trying to remove a category from Template:Skiing and Snowboarding-stub since I renamed the project. I can change the category on the page of the template, but when I go onto a page where the template is on, it shows the old category. What am I doing wrong? Kevin Rutherford (talk) 19:24, 28 February 2010 (UTC)

Probably nothing. Category changes in templates take a while to propagate through to articles. If it's not sorted itself after 24 hours, put another message here. --Redrose64 (talk) 19:37, 28 February 2010 (UTC)

More specifically, they take a while to go through the job queue. Graham87 07:39, 1 March 2010 (UTC)

I see they are now moving, albeit slowly. Looking deeper into this, I see that you made seven edits to Template:Skiing and Snowboarding-stub between 02:48, 27 February 2010 and 19:20, 28 February 2010. Of these, four of them changed the category in question (originally Category:Skiing stubs, it became Category:Skiing and snowboarding stubs, then Category:Barnstable, Massachusetts, then none, then Category:Skiing and snowboarding stubs). For each of those seven edits, all the pages transcluding the template will have had to have been placed in the job queue to be rebuilt. What I don't know is whether these requests are processed together (one job for seven changes) or in turn (seven jobs for one change each). --Redrose64 (talk) 17:50, 1 March 2010 (UTC)

Okay, that all makes sense now. At the rate it's going at, it will be emptied on Thursday afternoon. Kevin Rutherford (talk) 22:36, 1 March 2010 (UTC)

CustomizeGoogle

I need a new version of http://www.customizegoogle.com/ because Google's new layout doesn't have a Wikipedia link with that anymore.

Would someone who understands how to "convert the CustomizeGoogle extension to a zip file and then edit the files in that" please help me out? My friend says that's the way to get my beloved one-click Google-to-Wikipedia link back.

Anyone who knows how to do this, if you turn out to be local to me (and it's quite possible) then I will babysit your pets, kids, house, or fish as you see fit for a week. Thanks in advance!

Logoth (talk) 06:35, 1 March 2010 (UTC)

http://www.optimizegoogle.com/ took over the codebase. I've used it for a few weeks (it's identical to customizegoogle) and recommend it to all. :) -- Quiddity (talk) 07:28, 1 March 2010 (UTC)

How do you look at the source code? Logoth (talk) 03:49, 2 March 2010 (UTC)

Does that work with the new Google layout? They are asking for developers. Logoth (talk) 18:47, 1 March 2010 (UTC)

I made this ranked-choice voting poll so you can pick your favorite search engines to appear on the next version of OptimizeGoogle -- PLEASE VOTE WIKIPEDIA! :D Logoth (talk) 18:47, 1 March 2010 (UTC)

There is also a SourceForge project and a Google Group. Logoth (talk) 03:46, 2 March 2010 (UTC)

Manually adding talk page section produces wrong edit summary

When a user adds a talk page section via the "new section" link, the new section name appears in the C-style comment prefacing the edit summary. However, if the user manually adds the new section (by using "edit page" and typing "=="), the preceding section name appears instead. Presumably this is because, logically, it's text added to that section (although that interpretation could easily be repaired by detecting the "=="). The resulting incorrect edit history display is a source of minor inconvenience and confusion. Can we do anything about it, or is it a bug in the MediaWiki software which we don't have the power to fix? PL290 (talk) 13:34, 1 March 2010 (UTC)

The MW software sees the section number. For example, [3] pints to this section. See the part &section=31? That's what the software recognizes, and puts it in the edit summary when the page loads. This means that the edit summary is already filled in before you touch the editbox, and it is submitted alongwith the edit, not after. The software cannot tell if you added a section because the summary loads before you edit. You have to manually change the text between the C comments (case sensitive). You could make a script for that, which checks if there are extra section headings on a page with &section=... in the title, and changes the summary comments. ManishEarth^{Talk • Stalk} 13:47, 1 March 2010 (UTC)

Thanks. I realized my description was inaccurate anyway; it can't be with "edit page", since that generates no such preface, so the user in this scenario must be editing the last section. And as you say, the comment's already generated before anything's typed. Mind you, it's always struck me as odd that that's an editable part of the edit summary; were it generated separately instead, after Save is hit, that would seem more controlled, and in that case there would indeed be opportunity to interpret whether a new section's been added. I don't really want to make a script; more to make the right thing happen whenever users add new sections in this way. PL290 (talk) 14:00, 1 March 2010 (UTC)

Yeah. It's kinda weird that you can edit it. I tried it out at the sandbox. If you see this diff, you can see that I edited a section called "section test", which doesn't exist in both that revision and the previous revision. See the history of this page, too, the edit which I am making now has an edited section title. 59.183.154.54 (talk) 14:07, 1 March 2010 (UTC)

That's me above... Forgot 2 login. ManishEarth^{Talk • Stalk} 14:41, 1 March 2010 (UTC)

ImageAnnotator gadget on Wikipedia

I would like to propose for ImageAnnotator gadget to be installed on on Wikipedia. It was already discussed once discussed once here. ImageAnnotator gadget is a very popular gadget at Wikimedia commons where it is allowing users to provide more details to specific parts of the image. For example: name objects (and provide links) in panoramas, identify people in group images or highlight hard to see important details of the images. At the moment over 14k images on commons are annotated, for example this one. Please notice annotations identifying people available at Commons but lost on Wikipedia page. I would like to propose to install ImageAnnotator gadget on Wikipedia initially in the File: namespace only to allow more people get familiar with it. That would have an effect of making pages in file namespace on Commons and Wikipedia to carry the same information. --Jarekt (talk) 14:53, 1 March 2010 (UTC)

I have one big problem with this gadget, and that is the enormous amount of JS that it requires. To have such a tool enabled by default and distributed on all pages... I can already hear David complaining with his dialup connection and old computer. —TheDJ (talk • contribs) 16:37, 1 March 2010 (UTC)

I think it is possible to disable the gadget to individual people. --Jarekt (talk) 20:40, 1 March 2010 (UTC)

Oh definitely, but people who actively disable something are not our biggest audience :D It's just something that has me worried. Are there other wiki's besides Commons that have already deployed image-annotator ? —TheDJ (talk • contribs) 21:17, 1 March 2010 (UTC)

Personally, I don't find it especially useful on Commons, and it seems like it would be even less useful here. Mr.Z-man 22:11, 1 March 2010 (UTC)

As I posted in the last thread, my experience is that a substantial portion of "annotations" are useless test edits or "hi mom" stuff that doesn't get cleaned up - and of course if those are loaded here there's no immediate way to clean them up from this wiki. Another point that's been made is that Commons is not an encyclopedia and has only a limited NPOV policy, and some of the annotations are not really appropriate to display here for those reasons. If those hurdles could be overcome, it might be worth it, but I don't see that there's been any attempt to do so. — Gavia immer (talk) 23:30, 1 March 2010 (UTC)

Having said that, we could make the gadget available in the prefs for those who want it of course... —TheDJ (talk • contribs) 00:37, 2 March 2010 (UTC)

I'm not comfortable with automatically importing annotations from commons, because annotation changes make in commons won't show up on wikipedia article editors' watchlists. Identifying people in pictures is a potential BLP issue (e.g. imagine a vandal adding names of living people to pictures like File:Bloody Sunday-Alabama police attack.jpeg). Anyway, the popups could be a magnet for spamming external links onto images.

Technically I quite like parts of it, e.g. using frameless images and animations as popups. But we don't generally use large images or panoramas in wikipedia articles, and the only mainstream use I can see is if we decide to have a new type of encyclopedia page that is principally an image. Because that sort of page would be maintained on wikipedia (rather than on commons) all the usual watchlisting, NPP, neutrality and verifiability processes would limit the potential for problems. That model could be helpful for annotating details in large images such as diagrams, maps, aerial views (example) and maybe group photographs. - Pointillist (talk) 02:57, 2 March 2010 (UTC)

WikiMiniAtlas

Going off on a tangent, why is there no easy way to opt out of WikiMiniAtlas? As far as I can see (using IE8), WMA seems to slow down page rendering significantly, even for pages with just one coordinate; and makes it unworkable to include geolinks at all in long coordinate lists; yet the open source maps are currently inadequate for many purposes so it is usually better to click through to external maps. Is it true that there is no WMA opt-out parameter in coord templates (and also no opt-out in user-preferences)? And if there isn't, is there a reason for this? — Richardguk (talk) 02:13, 2 March 2010 (UTC)

Safari sort issue

See here for the question. I appended it to an earlier question, but it looks like I need to point it out again. Thanks.—NMajdan•talk 16:38, 1 March 2010 (UTC)

You should give more detail. What exactly goes wrong, and which exact column? It's unlikely that Safari would behave differently on different platforms here. —Simetrical (talk • contribs) 17:59, 1 March 2010 (UTC)

I have linked to the FLC where the exact issue is described.—NMajdan•talk 21:57, 1 March 2010 (UTC)

I'll just move the question here:

I have another sort issue, so I'll ask here rather than starting a new thread. A reviewer for an FLC I have up is saying the the postseason columns at List of Texas A&M Aggies head football coaches are not sorting correctly. I've looking in Safari on Windows, and it appears fine. Any help would be appreciated.—NMajdan•talk 16:52, 26 February 2010 (UTC)

I don't have any issues with Safari 4 here either. Unless the editor has a more accurate problem description (browser version, path of actions, screenshot of result), I don't think anyone can do anything else. —TheDJ (talk • contribs) 00:35, 2 March 2010 (UTC)

Invictus (film)

Hi everyone. I wanted to tell you that I made a mistake by renaming this article. I thought I was on the French WP page so I wanted to follow the French WP standards for films! Sorry for that. --TwøWiñgš Talk to me 22:56, 1 March 2010 (UTC)

I've moved it back to the proper format. — Gavia immer (talk) 23:22, 1 March 2010 (UTC)