Wikipedia:Village pump (technical)/Archive 52

From Wikipedia, the free encyclopedia
Jump to: navigation, search


Any way to watch a range of IPs?[edit]

My guess is that the answer to this is no, but I thought I'd ask on the off-chance: is there any way to track the edits of a range of IPs? For example, could I track all edits made by anyone in the 123.456.X.X range (that I just made up for this example)? TalkIslander 15:05, 30 November 2008 (UTC)

You can set a user interface gadget in Preferences / Gadgets that allows to look at Special:contributions,e.g via entering 123.456.0.0/16. --Tikiwont (talk) 16:59, 30 November 2008 (UTC)
That's certainly better than nothing - cheers! TalkIslander 18:26, 30 November 2008 (UTC)
Is there any way to sort by date rather than by particular IP? Tracking edits made by someone in a particular IP range would be a lot easier. Tejoka (talk) 02:19, 2 December 2008 (UTC)
There's a separate tool named anonEdits you can get from User:Franamax. It runs more slowly, but it's more flexible in terms of what it looks for. Since it works by examining recent change logs, it's more effective for looking for pattern vandalism from a dynamic IP range.—Kww(talk) 02:29, 2 December 2008 (UTC)
It isn't feasable to sort by date (server-side or in a javascript gadget). It is a limitation of the way the data is stored and retrieved. In order to accurately mix/sort by date, you must get _all_ contributions from all involved users (which involves a prefix search for a /16 or /24, or multiple username parameters for a /25 to /31). While, as-is, this gadget will eventually do it (it stops after 10,000 and asks you to make really sure you want to continue), I didn't feel actually having a 'sort by date' would reflect reality in case the queries aborted. You'd have misleading results due to missing data. Also, sorting thousands (potentially millions if you did a crazy /8) of table or li objects in javascript would probably assplode browsers. 10,000 almost crashes mine (locks up for a few minutes). I've pondered a feature letting you search multiple users (or a range) in a given narrow date range, but haven't gotten around to figuring out the UI or sort for it.
It is slightly feasable to sort the recent changes table data by date (server-side) since it is by default, but you'd have to make a regexy filter for it, and that would be a large amount of revisions to parse. Also, the data can only go back 30 days. --Splarka (rant) 08:27, 2 December 2008 (UTC)
Oh, I do have a tool that will show you a mixed 'recent changes' for multiple users: User:Splarka/contribsmulti.js. This is accurate and trustable for a given range, as it fetches 500 of each user's contribs, intermixes and sorts them by date, and only shows only 500 total. However, you cannot really paginate this data (there is no parameter for "give me N edits total for all these users", per above) and a limit of 5 usernames is a bit narrow for a CIDR. --Splarka (rant) 08:33, 2 December 2008 (UTC)

Autocomplete in the search box - a really cool feature, but...[edit]

I really love the autocomplete function added to the search box recently (well, maybe not that recently now...), but one objection I have to the way it works is that when you right click on a popup option it doesn't let you do anything (e.g. open in new tab, which would surely be helpful), but opens it just as if you had left-clicked it. Richard001 (talk) 00:47, 2 December 2008 (UTC)

I usually get around that by selecting (with the arrow keys) the one I want, then clicking off the link, then holding command (since I'm on a Mac) when I hit the Go button, which opens it in a new tab. I don't think there's a way to get the feature to work like a "regular" link; I've seen similar features all over, and that's how they all tend to behave. EVula // talk // // 17:14, 2 December 2008 (UTC)
The following simple script should patch the mwsuggest.js to open results in new window when you click while holding Shift. It also tries to do it on left and middle click but that doesn't work well in many browsers. —AlexSm 19:49, 2 December 2008 (UTC)
if (window.os_eventMouseup){
 var os_eventMouseup_old = os_eventMouseup
 os_eventMouseup = function(srcId,e){
  e = e || window.event
  var r
  if (e.shiftKey || (e.button>0)) r = os_map[srcId]
  if (r){
    var frm = document.getElementById(r.searchform)
    var inp = document.getElementById(r.searchbox)
    var oldT =, oldV = inp.value = '_blank'
  if (r){ = ''; inp.value = oldV }  

I don't follow so well... is that something I can add to my .js page here? I would really prefer a new tab than a new window too, having lots of windows isn't really my style of browsing (I suppose it's quite similar, though I can't move the windows around on the taskbar (possibly because it was designed by MS) and I can't go to 'undo close window' or 'recently closed windows). Is it possible for a higher level technical fix for this, e.g. changing the MediaWiki software behind it? Richard001 (talk) 03:25, 3 December 2008 (UTC)

In JavaScript, "new window" actually means "new tab or window, depending on your browser settings". Yes, the code is meant for your monobook.js. I guess we could open a bugzilla request for adding this Shift feature for everybody... —AlexSm 04:35, 3 December 2008 (UTC)

Monobook code problems[edit]

I inserted the the code importScript("User:Smith609/toolbox.js"); onto my monobook page, which was supposed to activate the citation widget in my browser, but it does not. I have reported this issue with the script creator, User:Smith609, but he was unable to resolve my problem. --Crackthewhip775 (talk) 06:02, 2 December 2008 (UTC)

What browser are you using? What javascript-related addons and settings are you using? Are you getting any javascript errors? Are you blocking javascript from any wikipedia-related site, of which the least obvious is Algebraist 08:57, 2 December 2008 (UTC)
Firefox, I have Java Quick Starter 1.0, and the the site you referred me to says my client is not allowed to access the requested object. I tried it on my other available browsers (Safari, Explorer) but it's the same problem. --Crackthewhip775 (talk) 23:36, 2 December 2008 (UTC)
Sure, you shouldn't be able to access that exact url, but blocking javascript from that source is a good non-obvious way of breaking custom js on wikipedia. Algebraist 01:11, 3 December 2008 (UTC)
It works for me, all it does is adding a new section with reference formatting links to the bottom of the left toolbar. Cacycle (talk) 01:21, 3 December 2008 (UTC)

Block template bug[edit]

the blocking template sometimes does not include </div> in the end so that the colour of the template to stop. see [1] > [2] if you know the source of the template, tell them... --CuteHappyBrute (talk) 15:13, 2 December 2008 (UTC)

I suspect that particular example was from {{test5-n}}, which was deleted in May 2008; I base this on the unique image used, which is referenced in someone's old "list of block notices" page. If you have any examples added more recently, please provide them. Anomie 16:44, 2 December 2008 (UTC)
yes that's the erroneous block template: [3] from then the colour sticks on the next edits> [4]. i have seen it many times. it didn't bother me until the 3-4th time i saw it. but still it may be from the 2007 deleted template as you said. thanx. --CuteHappyBrute (talk) 20:48, 2 December 2008 (UTC)
the lacking of <div> closure exists in many old templates.[5] [6] they probably have seen the bug. it's no big deal anyway. --CuteHappyBrute (talk) 21:18, 2 December 2008 (UTC)

How do script changes reach users if they don't clear their cache?[edit]

How do script changes reach users if they don't clear their cache? Many people use my script by adding the following to their monobook: importScript('User:Lightmouse/monobook.js/script.js'); It allows them to delink dates, reformat dates to US or international format, and to add metric units. Sometimes I will update it but the update doesn't have any effect on a user on the other side of the planet. When I get such a report, I tell them to clear their cache and that usually works. However, I don't know how my changes reach them if they don't clear their cache. Please can somebody explain this to me. Lightmouse (talk) 10:09, 2 December 2008 (UTC)

Well, they don't. At least not until the cache expires. There is a trick however that you can apply where you add a "version" to the script that does a "selfcheck" if it is still the current version. This is used by the edittools for instance: MediaWiki:Common.js/edit.js and MediaWiki:Edittools.js. --TheDJ (talkcontribs) 11:57, 2 December 2008 (UTC)

That would be very useful. I took a look at those .js files but couldn't work out exactly which piece of code I should copy. Can you make a suggestion? Lightmouse (talk) 16:35, 2 December 2008 (UTC)

What happens in this specific case is that edit.js reads a class of a certain HTML element. This class it tries to find is "edittools-version-*", with the * being the revision of the current script version. It then passes this version number as a paramater of "nocache" in the URL that is uses to load the script. This guarantees that a "new" URL is used to load the script, and that the browser cache is bypassed. Where this version# is stored is irrelevant of course, as long as it is not cached. (you could use an api.php call to retrieve it from a certain wikipedia page in your userspace for instance). It's not ideal, because you have to query everytime you retrieve it. An alternative idea here, is to check the last edittime of the wikipedia page of the script trough an api.php query and use that to determine if a script might be outdated. --TheDJ (talkcontribs) 00:20, 3 December 2008 (UTC)
wikEd uses a slightly different autoupdate mechanism: The script checks the current version number via Ajax in the background every two days (the last check date is saved in a cookie). If a new version is available it immediately reloads the page with cache-bypassing (window.location.reload(true)). It is all in the subroutine WikEdAutoUpdate. Cacycle (talk) 01:10, 3 December 2008 (UTC)

Thanks. I have created User:Lightmouse/monobook.js/cacheupdatetest script.js and copied WikEdAutoUpdate into it. Presumably, it would be better if I replace 'WikEd' with a different name. Are there any other changes I need to make it work for my users? 09:30, 3 December 2008 (UTC)

Yes, you need more helper routines (cookie handling, Ajax) and remove the stuff that you do not need (Greasemonkey). You definitely have to rename everything in order to avoid interferences and you have to test it thoroughly. Cacycle (talk) 13:54, 3 December 2008 (UTC)

I have made some changes but I am way out of my skill zone and I can't see the way ahead. I may have to give up. Lightmouse (talk) 14:06, 3 December 2008 (UTC)

wasted wiki real estate[edit]

it may have been raised before but on a long article there's big blue nothing down the left column of the page. Why not use a floating frame so that the stuff always there on the left column at the top comes down as you browse. Mccready (talk) 00:11, 4 December 2008 (UTC)

Because I have gadgets and javascript that makes it bigger than my screen and I want to access the stuff lower down. There is some javascript that can make it float if you want it. Dendodge TalkContribs 00:14, 4 December 2008 (UTC)
thanks. i meant for all users as a default. it would improve WP. Mccready (talk) 00:16, 4 December 2008 (UTC)
Because floating sidebars are ugly? Seriously, look at other major websites, most of them have some sort of sidebar, but few or none have one that floats as you scroll down the page. I don't see why empty space there is bad, it keeps the page from being just a big wall of text. Mr.Z-man 02:52, 4 December 2008 (UTC)
You can float the sidebar in your preferences. It works for some browsers, doesn't work for others. --Carnildo (talk) 03:39, 4 December 2008 (UTC)
You can? I don't see an option to do that in my preferences. – ukexpat (talk) 16:12, 4 December 2008 (UTC)

Discrepancy in article counts on {{Wikification progress}}[edit]

Background: {{Wikification progress}} is a template that shows the backlog of articles that need to be wikified. It shows the number of articles in each of the monthly backlogs (e.g. "Category:Wikify by <month> <year>"), recently tagged articles (Category:Articles that need to be wikified), and the entire backlog (Category:All pages needing to be wikified). The number of articles in the entire backlog equals the sum of the monthly backlogs plus the recently tagged articles.

Problem: I've just checked the template, and as of 3 December 2008, 23:59 (UTC) the article count for the entire backlog is 15,422, while the sum of the constituent backlogs is a whopping 16,022. That's a discrepancy of (exactly) 600 articles!

Following previous discussion on a suspiciously similar problem, I suspect that there is something wrong with the PAGESINCATEGORY count that is displayed in the template for one or more of the categories (these numbers do equal the page counts displayed in the categories themselves, though).

Can someone please verify if my suspicions are true. If so, then what is the best way of displaying the actual number of articles in the category? If not, then what else can be causing such a large discrepancy (I've already checked every template I could think of that includes wikification categories)?

Cheers. Liveste (talkedits) 00:33, 4 December 2008 (UTC)

This is a known bug; the category counts in the database can be incorrect for whatever reason and there's currently no way to refresh them. It seems to especially be a problem with categories populated by templates. Mr.Z-man 03:47, 4 December 2008 (UTC)
I don't suppose there's a better way to keep track of the number of articles in categories, is there? Liveste (talkedits) 05:57, 4 December 2008 (UTC)
It's not a problem of redirects is it?Mccready (talk) 17:27, 4 December 2008 (UTC)
No, redirects are included in the category counts – they're just listed in italics. Liveste (talkedits) 03:50, 5 December 2008 (UTC)

Glitch in arabic/persian text[edit]

I'm using IE7, and I see overlapping text in the Khwarezm article.

Does anybody else see this? (talk) 20:28, 4 December 2008 (UTC)

Nope, not me. Maybe it's an isolated case, or is it? Hytar (talk) 22:14, 4 December 2008 (UTC)
If the user could make a screenshot and show us the problem that way, then that might be helpful --TheDJ (talkcontribs) 02:32, 5 December 2008 (UTC)

Interface changes?[edit]

Am I taking crazy pills or have many of the editing interfaces changed recently? For instance, when editing a page, the following bitey text appears:

You are also promising us that you wrote this yourself, or copied it from a public domain or similar free resource. DO NOT SUBMIT COPYRIGHTED WORK WITHOUT PERMISSION!

Similarly, clicking on a redlink now shows the text

You have followed a link to a page that does not exist yet.

To create the page, start typing in the box below (see the help page for more info). If you are here by mistake, click your browser's back button.

Did I miss a discussion somewhere? Skomorokh 21:51, 2 December 2008 (UTC)

Those messages have been there for as long as I can remember... AmiDaniel (talk) 21:55, 2 December 2008 (UTC)
Odd. I may have changed my Monobook.js file recently; could that be what's doing it? The old redlink page used to have wording similar to "Try searching for the title. To start the article click here", and the old editing screen definitely didn't have the bolded allcaps (cruise control for cool) above. Skomorokh 22:03, 2 December 2008 (UTC)
They changed maybe a year ago, I think? --NE2 00:35, 4 December 2008 (UTC)
The messages are periodically made more and more glaring in a rather haphazard attempt to try to get the attention of users -- Gurch (talk) 01:59, 6 December 2008 (UTC)


Hi, I have been asking here and there about some problems graphics in Timeline have. The graphics I am making in Spanish wikipedia do not show data. Some times they show the bars and other times not. I read this is caused by a bug, but I don´t understand any about that. Somebody could help me? Thanks to everybody. This is the graphic. (If you can not see it, you have to click on edit):

Thanks again!--Eliasjorge4 (talk) 04:37, 5 December 2008 (UTC)

See bug 16085. MER-C 12:39, 5 December 2008 (UTC)


All the old versions of this image return 403 (HTTP forbidden) errors. There are no corresponding entries in the deletion log. Seems to be an isolated occurrence - I poked around the same directory using Google Image Search and couldn't reproduce the problem. But is it? MER-C 12:36, 5 December 2008 (UTC)

Didn't we have an issue with missing images a couple of months back? I'd say that it's related. EVula // talk // // 21:06, 5 December 2008 (UTC)
That was a different problem. The missing images return code 404, not 403. --Carnildo (talk) 22:00, 5 December 2008 (UTC)
bugzilla:16553 -- Gurch (talk) 01:57, 6 December 2008 (UTC)

Major UK ISPs reduced to using 2 IP addresses[edit]

-> WP:AN#Major UK ISPs reduced to using 2 IP addresses -- Gurch (talk) 16:00, 5 December 2008 (UTC)

Mixed Content on Secure Site[edit]

I've noticed that when using the secure site (, there are images (the WP Egg, banners, etc) and other content that are linked from the regular non-SSL site. This causes a pop-up box in IE at every page load, stating that the page contains secure as well as insecure content, and do you want to display the non-secure content. In Firefox it throws a security warning icon, but at least doesn't force you to acknowledge a popup box at every page load. Would it be possible to address the http:// links from the secure site, so all content is loaded from the secure site? It would make using the secure site more appealing under the IE browser. ArakunemTalk 20:55, 5 December 2008 (UTC)

This will not happen any time soon. Primarily as I understood it, because then all images would need to be served trough https, for which the server capacity simply is not available to the WMF at this moment. --TheDJ (talkcontribs) 01:50, 6 December 2008 (UTC)

CSS font for zh-Latn[edit]

A standardised method exists for Romanising [Mandarin] Chinese text, known as Pinyin. This uses Latin characters with accents. Currently setting the language code using {{lang|zh-Latn|...}} for such text, may cause some browser combinations to only read as far as the zh (Chinese) and select a double-width font; if this happens the result is ugly. The solution is probably to override the font (back to the default) using CSS. Assistance and guidance on the changes needed would be welcomed. One error case noted by another editor at Template talk:Zh-p#zh-Latn is Firefox 3 on [a] WinXP/SP2 desktop. Leaving the text without a lang= tag means that the span is assumed to be English, which it is not. —Sladen (talk) 23:36, 5 December 2008 (UTC)

wikipedia suddenly in plaintext (?) mode?[edit]

I access wikipedia via my work machine, which is subject to all sorts of restrictions and can't access, and my home machine which isn't restricted. i never changed my skin, so it's still classic i guess. anyway, my home machine still shows it, but my work machine all of a sudden shows wikipedia in kind of a flat text mode; no bars across the top with the views or my personal tools, no bar down the left with the navigation links, search, interaction links, toobox; instead they are all tacked onto the bottom of the page in a vertical set of plain text links. also, the navigation popup which used to show the beginning of the page content and links to the various views on mouseover a link doesn't work on the work machine, but still works at home. so.. did the css or something get moved over to another domain, like mediwiki? is there any fix? thanks. Gzuckier (talk) 00:45, 6 December 2008 (UTC)

  1. reload (Ctrl + F5 or Ctrl + R as appropriate)
  2. clear your browser cache
  3. try this link
  4. check your user CSS pages
  5. check your browser font/color settings
  6. complain to your workplace / get another job somewhere that doesn't have stupidly restrictive Internet filtering

-- Gurch (talk) 01:54, 6 December 2008 (UTC)

I've had the same problem with IE6 at school, its only happened recently (I only noticed the problem on Friday). The best option would be to upgrade to a proper web browser, but since that isn't a option for some users I'm going to see if I can replicate the problem --Chris 12:19, 6 December 2008 (UTC)


I am trying to use the Template:GeoGroupTemplate and it works just fine on the article page with the coordinate listing. I asked a query on the talk page about using the list on one article, and the maps on a different article, and was replied to but I don't understand still. Usage of geogroup template on one page, listing on separate page The article North Sea is very long. I would like a geographical features map on the page using coordinates so the places are rather precise on the map of the North Sea, so the coordinates are all listed at Geography of the North Sea and the geogroup template there is awesome. Just need it at North Sea as well. Also asked at Wikipedia:WikiProjekt Georeferenzierung/Wikipedia-World/en Kind Regards SriMesh | talk 02:33, 6 December 2008 (UTC)

Yes check.svg DoneI have now added this template, User para helped out on the aforementioned talk page. SriMesh | talk 04:41, 7 December 2008 (UTC)

Help with database dumping?[edit]

I was told to come here after asking elsewhere. Apologies if I'm in the wrong place yet again. I was wondering if someone could help me, or point me to someone or somewhere that could, convert a database dump to multiple, non-treed base-raw html pages. I've already followed the somewhat bland instructions at Wikipedia:Database but never get quite the results I needed.
I have I want to convert the uncompressed code (using ANY programs) to multiple html pages. The result I'm looking for at the end is to have pages the equivalent of going to every page and clicking save:as... and just have a bunch of html files in a folder. Am I missing something? Lostinlodos (talk) 03:24, 6 December 2008 (UTC)

Those database dumps are in wiki code and not HTML - you're looking for something like (14 GB, example page). Unless you feel like splitting the dump into individual pages and running them through an offline version of the parser as required it looks like you're in for another long download. MER-C 09:43, 6 December 2008 (UTC)
Mer-c, is that going to give me the complete en wikipedia site in html form? I wouldn't mind another download if it's what I was looking for initially and saves me the time of converting code. Lostinlodos (talk) 18:56, 7 December 2008 (UTC)
Yeah, you'll get a big tarball of (HTML) pages like the one I linked to. The problem is that they are six months out of date. MER-C 01:03, 8 December 2008 (UTC)
Perhaps it would be useful if you gave some background what you're hoping to extract from the dump and for what purpose; and then we might be able to point you to the easiest/most suitable format. —Sladen (talk) 01:27, 8 December 2008 (UTC)
For lack of better way to put it, I plan to read, from a frozen snapshot, wikipedia. Page by page. Lostinlodos (talk) 04:45, 8 December 2008 (UTC)
Just in one language? That could take a while. At five-minutes per articles, it would take 24.5 years; if you were planning to sleep each night, it would could be 37 years—and with the weekends off—52 years! Did you have a specific area you wished to focus on? —Sladen (talk) 04:56, 8 December 2008 (UTC)
But it doesn't take five minutes to read a stub. ;-) Meanwhile, a featured article on a complex subject could take two hours to get through. And are we factoring in the distraction variable? :-) --MZMcBride (talk) 04:59, 8 December 2008 (UTC)
I guess this is not the right path either. I'm not looking for "stubs". I'm looking for creating a mirror in HTML format.
I'm looking for an alternative to: I type in and click ctr+s. If I were to type in SLI in the search box and then do a ctr+s. Then type in carpet and do a ctr+s, then type in pencil and do a ctr+s. To be clear, if wikipedia had only three articles, I would have downloaded wikipedia. I'd like to download 'something' and then end up with all of wikipedia's articles likewise in a single folder of many individual HTML pages. Does that make more sense? As for 5 minutes per article, I do read a bit faster than that. ;) Lostinlodos (talk) 14:59, 8 December 2008 (UTC)
To echo MER-C, it sounds like is what you're looking for, but to be honest, I'm not sure what use it really is. Its a 14 GB compressed file. I don't feel like downloading the entire thing to test, but a test with the much smaller Abkhaz Wikipedia gave nearly a 7.1-fold increase in size on decompression, so the full en.wikipedia one would be more than 98 GB, and the dump doesn't include images. Most computers capable of handling that probably have an internet connection. Mr.Z-man 02:16, 9 December 2008 (UTC)
Many thnks, It appears that is exactly what I'm looking for, no images, no nonsense, raw HTML that I can read as pure text files. Thanks everyone!Lostinlodos (talk) 02:57, 9 December 2008 (UTC)

What's up with the invisible graphic on the Wikipedia homepage? [SOLVED][edit]

There's a non-showing graphic (new, I guess — I've seen the little box for it for about a week) at the very top of the international Wikipedia home page. It doesn't show up on either of my browsers (AOL or IE6, on a PC). What's the problem, and when is it going to be fixed? Thanks. Softlavender (talk) 05:10, 7 December 2008 (UTC)

You mean It seems to be another 0 byte image. Algebraist 05:13, 7 December 2008 (UTC)
Yes, that's it. I can't even see it when I click that link. What's up with it, and when is it going to be fixed? Softlavender (talk) 05:20, 7 December 2008 (UTC)
Is that image one of those that got lost in that image server wipe-out a bit earlier? Or do we have a different problem? Calvin 1998 (t*c) 06:08, 7 December 2008 (UTC)
By the way, the problem is that the image is supposed to contain something, but for some reason it doesn't, probably because it got deleted server-side. Calvin 1998 (t*c) 06:09, 7 December 2008 (UTC)
Works for me, though this is what it is supposed to look like. MER-C 11:36, 7 December 2008 (UTC)
Works for me now as well... Calvin 1998 (t*c) 20:26, 7 December 2008 (UTC)
  • OK thanks, guys, looks like they fixed it (though the image now is different from what that invisible file actually contained!). Softlavender (talk) 01:07, 9 December 2008 (UTC)

Reviving topic (Optimize template code)[edit]

I'm here to mention an archived topic that nobody seemed to respond. It's about simplifying a template code, which has obvious repetitions. I wonder if that issue is a bit too "technical" for the guys here. Hytar (talk) 22:15, 7 December 2008 (UTC)

6 Tabs For All Projects[edit]

I don't care how the tabs are arranged, be they 3 by 2 or 6 by 1, but 6 tabs, instead of 4 needs to be implemented.

What I am talking about is that, say, for this page, we have "project page", "discussion", "edit this page", and "history", at the top. These are for navigational, editing, forensic, and learning purposes. Now, this isn't very user friendly, and people have argued that having 6 instead of the 4 we have now would make it more clumsy. And that if anons made an account, they could use CSS to implement this. Well some anons don't like making accounts. And it is more clumsy right now, and confusing, where by pressing the discussion tab, and then edit, it is counter-intuitive. People also lose thoughts, depending on their connection speed, by having to wait for extra pages to load. It's even hard to understand with the 4 tabs switch over system that is used right now.

Another problem are the templates: going to Template:Navbox Province of Italy, at its top left corner, you will see vde. This too needs the 3 other analogs. You can put them in 3 by 2, or 6 by 1, but for the same reasons above, specifically for server resource times/loading times and userfriendlyness/navigational ease. (talk) 12:46, 6 December 2008 (UTC)

These are two separate issues. The tabs at the top of the page are controlled by settings in the MediaWiki software. The vde links are within the {{Navbox Province of Italy}} template and are a function of the {{Navbox}} template it is based on; other templates may not have vde links. --—— Gadget850 (Ed) talk - 13:40, 6 December 2008 (UTC)
I really would never want someone to add a comment on the discussion page without seeing what was there already. As to losing your thoughts, keeping a notepad open for those works well. (talk) 16:37, 9 December 2008 (UTC)

Spaces in external links[edit]

At {{LGV Est}} there are a couple of external links that contain spaces; what can I do to make them render properly? ChrisDHDR 13:21, 7 December 2008 (UTC)

Try replacing the space with the entity %20 (example:
-=# Amos E Wolfe talk #=- 15:03, 7 December 2008 (UTC)
Thanks for the help though shouldn't this be put somewhere for people with the same problem? ChrisDHDR 15:35, 7 December 2008 (UTC)
It and other character conversions are at Help:URL. See Special:WhatLinksHere/Help:URL for pages which currently link to it. Maybe some others like Wikipedia:External links also should. PrimeHunter (talk) 16:21, 7 December 2008 (UTC)
The easiest way to convert links would be to copy and paste them from the browser URL field after you've visited the site. I don't know of a browser that leaves the space as-is. EVula // talk // // 20:15, 9 December 2008 (UTC)

Entering "wiki:" into search box redirects to[edit]

Something Rdunn (talk · contribs) discovered[7]: When you type "wiki:" into search box, it redirects to instead of showing the "bad title" message. Anyone have an idea why that happens? Oo Regards SoWhy 12:51, 9 December 2008 (UTC)

It's listed on meta:Interwiki map. PrimeHunter (talk) 13:00, 9 December 2008 (UTC)
I knew someone here was quick to know. Thanks for the info :-) SoWhy 13:09, 9 December 2008 (UTC)

Which articles are in a non-existent category?[edit]

It's easy to find out which pages link to a non-existent article, but how do I find out which pages are in a non-existent category? Specifically, I went to category:Coordinate geometry and clicked on "edit" and then on "what links here" and found nothing, although there was an article that bore that category tag. (The article was law of tangents, from which I removed the tag because it doesn't fit even if there were such a category.) Michael Hardy (talk) 23:45, 9 December 2008 (UTC)

...OK I figured it out: just use the "preview" button. Michael Hardy (talk) 23:51, 9 December 2008 (UTC)

Re. Wikipedia:Village_pump_(proposals)/Archive_38#Preview_comments[edit]

I'm creating a new thread for that topic because it died out back then, but no one objected either. Do people think those ideas should be implemented? I think it would improve preview, which is already a very useful tool. -- Mentisock 16:30, 2 December 2008 (UTC)

This is about (re-)moving the "This is only a preview page" message as it allegedly interferes with the article rendering. I do not yet see a need for it as templates are substituted for previews and we do not care about pixel-exact positioning. Maybe you could explain your problems in more detail.Cacycle (talk) 00:58, 3 December 2008 (UTC)
Well, it's an entire heading and plain text... it's not about being pixel-exact but it takes space in the actual page (which should be just about previewing what the changes will really look like, and as it is even previewing the page without changes renders a differently-styled page because of the notice) solely to remind the editor (redundant with so many reminders - at least shouldn't there be an option to turn it off?). And about the templates - substed templates aren't rendered in the previewed edit box (like I can't just try to {{subst:test}}, preview it and then be able to preview the substituted code as well). Although I suppose there might be some technical restrictions on the latter one. -- Mentisock 17:42, 3 December 2008 (UTC)
{{subst:test}} expands as expected for me and I still do not see your problem - or is this a joke? Cacycle (talk) 03:22, 4 December 2008 (UTC)
That template would only expand upon saving the page, thus necessitating test edits for experimentation. And... how possibly could this be a joke? O_o A minor problem maybe, but why would I post here if it wasn't real? -- Mentisock 13:25, 6 December 2008 (UTC)
They seem to expand on preview for everybody else... DoNotFeedTroll.svg Cacycle (talk) 01:59, 8 December 2008 (UTC)
[8]. -- Mentisock 09:06, 8 December 2008 (UTC)

I see nothing wrong with the current preview. WP is not a text editor, it is an encyclopedia. If after saving you don't like what you get, just go to history and hit undo. (talk) 16:27, 9 December 2008 (UTC)

Neither do I, but then again, it works for me. – Alex43223 T | C | E 02:23, 11 December 2008 (UTC)

Use of external data for internal calculations[edit]

In the discussion at Rupee-dollar conversion, I proposed copying the U.S. Government historical tables for the rupee-dollar conversion multiplier to be used in a monetary conversion template. However, it would be easier if Wikipedia could tap into free external conversion tables (such as available at U.S. Federal Reserve) for internal conversion template usage. Is there a way to do this with Wikipedia software? -- Suntag 14:16, 5 December 2008 (UTC)

What, you want to put all us editors out of work? What would we do for income? (talk) 16:31, 9 December 2008 (UTC)
Apparently we'd convert money for profit :-) Nyttend (talk) 23:03, 10 December 2008 (UTC)

Own Wiki[edit]

I would like to set up my own Wiki site (not to create an encyclopedia!) and wonder what is the best way. I would anticipate that, at most, it would have 1,000 contributors and perhaps 50 concurrent users. Should I have a company host the MediaWiki software for me? Is there a company that allows you to "spawn" a whole new MediaWiki installation? Any advice on the best way to go, would be appreciated. Twotinsofbeans (talk) 16:08, 6 December 2008 (UTC)

This page is for technical questions about Wikipedia. Try Wikipedia:Reference desk/Computing. Algebraist 16:10, 6 December 2008 (UTC)
If you want to run your own MediaWiki-based wiki, you (or someone who works for you or on your behalf, such as a system administrator) should read everything linked under WP:EIW#MediaWiki. The more you know about what you're doing before you start, the easier it will be. Setting up your own wiki, and making it work, is a nontrivial task. It's relatively easy to download and install MediaWiki, but what you get is a very bare-bones wiki, which lacks many features you will sorely miss if you have much experience editing on a well-developed wiki such as Wikipedia. So you have to learn about how to port features from Wikipedia to your own wiki. Then the software is only half the challenge - the other half is to build your user community. Wikipedia is by comparison an online paradise because here we have thousands of skilled, motivated, and benevolent users. When you start a new wiki from scratch, you face a chicken and egg problem - your wiki will suck until you have an established community of expert users to make everything nice, but until you get that community, few people will want to join your wiki (because it will suck). It's very hard for a new wiki to get over that initial hurdle. Obviously it helps if you already have a few dozen expert wiki users who want to join your project, but if you did, then I suspect you would not need to ask here for advice. But don't let this discourage you, lots of people have faced these problems and overcome them. You just have to be smart and determined. Perhaps the easiest way to get started experimenting on your own is to set up a Wiki on a stick, for which you only need your own computer, without the complications of working on a server. (You can also use the wiki on a stick method to make local working backup copies of your wiki, when you get it running on a server. This protects you and your users against catastrophic loss of data.) This type of question comes up on the Help desk from time to time; you can read some previous answers with this search:
--Teratornis (talk) 18:49, 6 December 2008 (UTC)
I notice that your contributions under this account are not extensive. If you don't have much editing experience on Wikipedia, you might consider gaining experience here before starting your own wiki. The Wikipedia community has worked out the whole process of how to build a wiki better than almost anyone else. If you accumulate a few thousand edits here, you will have absorbed many subtle lessons that you will need when you run your own wiki, and on your own wiki there might not be anyone there to teach you. It would be much harder to reinvent on your own all the wheels that Wikipedia has already invented. --Teratornis (talk) 18:56, 6 December 2008 (UTC)
I'm with you Twotinsofbeans I like Wikipedia, and helpful people like Teratornis; but I also like the feeling of not having people hovering over my shoulder, waiting to huff an article the moment I create it, or argue and argue over some little edit or link. You can learn by following examples and getting tips, like the useful ones here, but you also learn by doing. Good luck on your Wiki and your time here. :-) Yartett (talk) 01:03, 11 December 2008 (UTC)
Have you tried requesting a new wikia at ? ~AH1(TCU)

CSS for lists without bullets[edit]

Templates like {{Mozart operas}}, which currently use <br /> to separate list items, should be marked up as proper lists, with the bullet styling removed to preserve the current appearance.

I've created a sandbox version at User:Pigsonthewing/scratchpad, using list-style-type:none on the parent wrapper, but that doesn't work and I can't figure out how to use styling in wiki mark-up to remove bullets. Can anyone assist, please? (Feel free to edit my scratchpad page.)

Once the style is perfected, I intend to ask for it to be moved to the main stylesheet, and applied with a class. Andy Mabbett (User:Pigsonthewing); Andy's talk; Andy's edits 00:29, 8 December 2008 (UTC)

You can't really do it with wikimarkup, as the square bullets come from a style applied to the <ul> rather than the table. So to do it, you'd have to make the table with HTML:
<ul style="list-style-type:none; list-style-image:none">
<li>''[[Die Schuldigkeit des ersten Gebotes]]'' (1767)</li>
<li>''[[Apollo et Hyacinthus]]'' (1767)</li>
<li>''[[Bastien und Bastienne]]'' (1768)</li>
--Mr.Z-man 01:14, 8 December 2008 (UTC)
Thank you, but there is definitely a way to do it. It's done on {{flatlist}}, but I can't recall, or work out, how. Andy Mabbett (User:Pigsonthewing); Andy's talk; Andy's edits 13:04, 8 December 2008 (UTC)
{{flatlist}} uses class="horizontal". If you want to go that route, create a new class and propose it gets added to MediaWiki:common.css. Chris Cunningham (not at work) - talk 13:31, 8 December 2008 (UTC)
You could also simply use ":" (which outputs a defenition list using <dd>):
Die Schuldigkeit des ersten Gebotes (1767)
Apollo et Hyacinthus (1767)
Bastien und Bastienne (1768)
Only drawback is the slight indent. EdokterTalk 21:53, 10 December 2008 (UTC)

Search results "stuck"[edit]

I have put some time into removing every instance of the string "United World Chart" from Wikipedia, and, so far as I can tell, there is only one occurrence left (legitimate, because it's in quoted text inside a reference). I got rid of the remaining 1200 last week. However, when I use MediaWiki search to look for "United World Chart" to make sure people don't add references back in, I consistently get 40 results, day after day. How long will it take for the search function to catch up to reality? On the flip side, how long between the time that someone adds it back and it showing up in the search?—Kww(talk) 03:58, 9 December 2008 (UTC)

Good question, same happens for me. How long does it take? – Alex43223 T | C | E 09:09, 9 December 2008 (UTC)
As far as I can see, all of these are valid hits. Remember that the search searches the wikitext not html. So if you look at the wikitext, you'll see "United World Chart" as a template parameter. The index is updated daily (early morning GMT), and you can see the date of the indexed page as Date: on the search results. --rainman (talk) 11:03, 9 December 2008 (UTC)
I'd swear I checked for that five times, but there it is, plain as day. Thanks.—Kww(talk) 11:17, 9 December 2008 (UTC)

Search results "stuck" redux[edit]

Worked for 39 out of 40 articles. However, a search for "United World Chart" in quotes still returns World as a result, and I cannot see why.—Kww(talk) 12:55, 10 December 2008 (UTC)

You mean UWC? Some data (like links to the page) are cached with the article in the index for performance. I've made a null edit to this article, that should clear it in next index update. --rainman (talk) 18:25, 10 December 2008 (UTC)

Section links in edit summaries[edit]

I'm sure this has probably been discussed before somewhere, but would it be possible for the section title that is automatically generated in an edit summary to be a link to that section header? ~ JohnnyMrNinja 21:06, 9 December 2008 (UTC)

What's wrong with clicking on the little arrow thing? Algebraist 21:08, 9 December 2008 (UTC)
It's a little indistinct. I assume JohnnyMrNinja had just never known about it, as I didn't for a good month or so. Chris Cunningham (not at work) - talk 21:10, 9 December 2008 (UTC)
Thanks for the information. I have over 8,000 edits and I never noticed that the arrow was actually a link. Dbiel (Talk) 23:01, 9 December 2008 (UTC)
Oh, hey, there's an arrow there... I guess that answers what's wrong with it. I know most people just edit from their PC, but I actually can't see that character from my smartphone. Linking the text instead of a non-standard character might make it useful for everyone. I don't know how it does with a screen-reader. Also, I just think that if the text were linked it would be much more intuitive. ~ JohnnyMrNinja 19:23, 10 December 2008 (UTC)

"New external links" that do not exist[edit]

When I edit Wikipedia articles, I often do it from a text-only environment using lynx. Because I do not have an account with Wikipedia (and have no intention of having one), this prevents me from posting any new external links because I cannot pass the anti-spam captcha test.

I have no problem with this, but I do get annoyed when I am told that my edit includes new external links but it does not.

Just now I tried to add a response to the current version of a Reference Desk item, WP:RD/E#NFL playoffs, and teams with 2 ties, and was incorrectly told that my edit included new external links. I then backed up, edited the item again, and tried saving it without adding or changing any content -- and I was still told that my edit included new external links.

What is going on here and can it be fixed?

In case it matters, the current oldid for WP:RD/E as I write this is 256982623; of course I can't post a direct link to that version of the page because a URL would be treated as an external link.

(Please direct any replies here and not to the talk page for my IP address, which is shared.)

-- (talk) 06:06, 10 December 2008 (UTC)

It just let me make an IP edit to the page without bringing up any kind of WP:EL warning. (Mac OSX + Safari.) Are you in a position to try making the same edit both with Lynx and with MSIE or similar and see if you get the same message with both? If not, we'll know it's a bug in Lynx and not in the MediaWiki software. – iridescent 16:17, 10 December 2008 (UTC)

Tracing templates from categories[edit]

Is there possibly no way to find out which templates (transclusions utilizing includeonly) are populating a particular category? -- Mentisock 09:44, 10 December 2008 (UTC)

How does one influence the height of cells in tables?[edit]

Is it possible to use height the same way as one can use width to control the size of cells in tables?

For example, when using sub tables in a parent table, setting the sub table width to 100% will cause the sub table to be "evenly" spread across the cell in the parent table. However, setting height to 100% does not cause the sub table to be spread in any way up and down the cell in the parent table.

Setting the height of the sub table to a set number of pixels does cause the sub table to be a minimum height, but this does not cause the rows in the sub table to be spread at all over the full height.

Setting the height to a percentage does not seem to work at all, using either the deprecated height="100%" or using style="height:100%;".

Any ideas on how to get a table's rows to be spread evenly vertically over the available height?

Peet Ern (talk) 10:29, 10 December 2008 (UTC)

Pop-ups not working?[edit]

Have pop-ups stopped working for anyone else, or just for me? DuncanHill (talk) 23:06, 10 December 2008 (UTC)

Fine here. Algebraist 23:09, 10 December 2008 (UTC)
Hmmm, everything went painfully slow for a while, so I closed my browser and opened it again, and now it seems back to normal. DuncanHill (talk) 23:13, 10 December 2008 (UTC)
Happens to me sometimes, and that always seems to work. – Alex43223 T | C | E 06:22, 11 December 2008 (UTC)

Delete not deleting?[edit]

An IP just added a large "History" section to Mountainair, New Mexico, which I've reverted because it's a copyvio of Mountainair's website. Being an administrator, I decided to delete the revision (can't hurt), so I deleted the article and proceeded to restore it: before restoring I clicked just the copyvio revision to restore and then selected "Invert selection", so I know that I told it to restore all revisions except the copyvio. However, upon restoring it, I see the edit still there (done at 22:48, 10 December 2008 by, so I ended up reverting it instead. Can anyone explain why the edit didn't stay deleted, even though I told it to restore only the other edits? Nyttend (talk) 23:09, 10 December 2008 (UTC)

I was going to say that sometimes you need to wait a few minutes for everything to look right after deleting and restoring revisions (especially when merging histories), but I don't think that is the case here. I tried to do exactly what you said you did (invert selection and all that), and now it appears what I did has deleted the one copyvio edit. I cannot say why mine worked and yours didn't. Are you positive that one box was unchecked? The odd thing is that the deletion log says we both restored 39 edits. I can verify that yours did not work because before I touched anything, I did not get the "View or restore one deleted edit?" message on the history page (but now we do). Hmmm... anyone else with ideas? -Andrew c [talk] 00:11, 11 December 2008 (UTC)
Yeah, I'm certain that I checked the box. The answer to the 39 is that you undeleted my reversion of the copyvio. What I did was click the box, hit invert, type my rationale for restoring "Restoring minus copyvio", and hit Tab and then the spacebar. After a minute, I realised that nothing was happening (apparently spacebar doesn't select the restore button) and clicked the button, whereupon it restored the page. Perhaps I wasn't paying attention and accidentally selected the copyvio edit with the spacebar? Nyttend (talk) 00:20, 11 December 2008 (UTC)
No, this isn't it. Check the edit history for User:Nyttend/sssss, which I just created and deleted according to the process I followed with Mountainair; and the spacebar didn't work this way. Nyttend (talk) 00:24, 11 December 2008 (UTC)

Watchlist to show most recent edit tag - "top"[edit]

Is there any way for the watchlist to add the "top" tag as is done on the other similar lists such as "my contributions" and "user contributions" It would be a great help to see at a glance that the edit was the most recient edit on a page


Dbiel (Talk) 19:17, 9 December 2008 (UTC)

It only shows top edits anyway. They may not be top by the time you get to them, but since the watchlist is static until you refresh, there's nothing it can do about that. ♫ Melodia Chaconne ♫ (talk) 20:04, 9 December 2008 (UTC)
The static point is no different that the other lists ie "user contributions". That is not a problem, as you said, all you need to do is refresh. I just do not see what would be so difficult to make the watchlist function the same as the user contributions list. Dbiel (Talk) 22:57, 9 December 2008 (UTC)
What exactly are you asking for? Your watchlist will show ONLY top edits, which isn't the case in the 'user contributions' list. ♫ Melodia Chaconne ♫ (talk) 23:19, 9 December 2008 (UTC)
What are you talking about? I will outdent and post a small section from my watchlist
It definately shows a lot more that just the top edits, which I happen to like. It just does not include the top tag on the most recent edit, which is what I would like to see changed. Dbiel (Talk) 23:58, 9 December 2008 (UTC)
This is only if you have "Expand watchlist to show all applicable changes" enabled in your Wikipedia preferences (Watchlist section) --TheDJ (talkcontribs) 00:53, 10 December 2008 (UTC)
Yes, I am aware of the "Expand watchlist to show all applicable changes". What I would like to see is the (top) [rollback] which appears on the user contributions" list, but not on the watchlist. Dbiel (Talk) 01:08, 10 December 2008 (UTC)
But you failed to inform the others of having enabled this, which confused them, because they were assuming the default behaviour, in which the (top) marker makes not sense. --TheDJ (talkcontribs) 03:42, 10 December 2008 (UTC)
Sorry for the confusion; so I guess the correct question would then be: With Expand watchlist to show all applicable changes enabled, is there any way to display the (top) marker (tag) to identify the most recent edit? It would then be consistant with the "user contributions" page Dbiel (Talk) 05:58, 10 December 2008 (UTC)

Something that I just discovered, and felt should be noted here is that this request was actually posted over 1-1/2 years ago and remains as a new feature request to the WikiMedia Software. see Dbiel (Talk) 03:16, 12 December 2008 (UTC)

Color of wikilinks[edit]

I asked this question mid-November at Wikipedia talk:Colours but to no response. Maybe it will get better reception here: What are the hex codes for the colors shown in Wikipedia internal links?
-What color is the link to this unvisited page?
-What color is the link to this non-exsistant page?
I noticed these colors are different from <font color="blue"> and <font color="red">. Also, what are the colors to visited, existing pages (they're more of a purple color)? Thanks for your help, Arbitrarily0 (talk) 03:01, 10 December 2008 (UTC)

Somewhat skin dependant but some of the defaults are:
A standard wiki link you have not visited yet = 002bb8
A standard wiki link you have visited = 5a3696
A non-exsistant page you have visited = a55858
An external link = 36b
See Wikipedia:Catalogue of CSS classes for links to all the information And more specificly to the default skin see which was used to find the above color codes Dbiel (Talk) 05:49, 10 December 2008 (UTC)
Thanks Dbiel! I placed a copy of your note back at Wikipedia talk:Colours. Arbitrarily0 (talk) 21:17, 10 December 2008 (UTC)
And we have a help page at Wikipedia:Link color, which could use some work. --—— Gadget850 (Ed) talk - 20:36, 11 December 2008 (UTC)

Thumbnail Generation Broken[edit]

Derwent river as seen from Poimenna Reserve, Austins Ferry

All of a sudden thumbnail generation for the attached image is broken. This isn't good etc. Any suggestions for a fix? Noodle snacks (talk) 12:20, 10 December 2008 (UTC)

The image is 16.7 MB and and 12919x1892 pixels: quite a lot to process whenever one just needs a thumbnail :) The error message given is:
Error creating thumbnail: convert: Insufficient memory (case 4) `/mnt/upload3/wikipedia/commons/b/b9/Austins_Ferry_and_Derwent_River_from_Poimenna_Reserve.jpg'.
convert: missing an image filename `/mnt/upload3/wikipedia/commons/thumb/b/b9/Austins_Ferry_and_Derwent_River_from_Poimenna_Reserve.jpg/250px-Austins_Ferry_and_Derwent_River_from_Poimenna_Reserve.jpg'
So I uploaded a new one several times smaller, Image:Austins Ferry and Derwent River from Poimenna Reserve 1292.jpg, which is 1800x264 and 391 KB, and which MediaWiki seems to have no problem making a thumbnail of. You can use the resized version instead of Image:Austins Ferry and Derwent River from Poimenna Reserve.jpg in articles, since the former points to the latter in the image description, and thumbnail generation of the resized version is a lot less likely to fail. Alternatively, because the quality of the 1800x264 version seems poor, feel free to upload a somewhat larger resized version. GracenotesT § 14:35, 10 December 2008 (UTC)
I am not keen to replace it with a resized version in the articles as there would be quality loss, and that version was voted as a featured picture. Has there been some server setting change to cause the problem? Noodle snacks (talk) 07:41, 11 December 2008 (UTC)
Mm, I'm not quite sure if any server settings were altered or if it's just a temporary hiccup. Consider, though, that there is also an inevitable significant quality loss implied in making a 16.7 MB image into a thumbnail. Whether the server generates a resized version or one is pre-uploaded, the quality reduction is the same; the only difference is that the latter is a lot less hard on the server, and rendering it is less likely to fail.
Wikipedia:Featured picture candidates/Internet Map is an example of a featured image, a diagram, that has a larger version. With panoramas/photographs, it's a bit different, but the basic idea of having a smaller version that's easier to render/deliver to the article viewer is the same. GracenotesT § 22:04, 11 December 2008 (UTC)

Intrusion attempts on edit pages?[edit]

Resolved: Run liveupdate to download the latest updates.

Starting today, whenever I attempt to edit a page on English Wikipedia, I get an intrusion attempt alert from Norton Internet Security. (Commons and German Wikipedia appear unaffected.) The name of the "risk" is "HTTP Acrobat PDF Suspicious File Download". I suspect it's an error in the risk definition, causing it to by chance match something on the edit page, but I can't be sure. Nothing obvious on the edit page fails to work. But it sure is annoying. Any ideas? Powers T 14:46, 10 December 2008 (UTC)

See also Wikipedia:Administrators'_noticeboard#Wikipedia_triggering_Norton_Internet_Security_alerts --TheDJ (talkcontribs) 15:20, 10 December 2008 (UTC)
Not encountering this with my system, but we should get this concentrated in a single thread. Hiberniantears (talk) 15:36, 10 December 2008 (UTC)
All relevant threads have now been pointed to this thread. seicer | talk | contribs 15:42, 10 December 2008 (UTC)
Thanks. I took no action hoping someone would make a sounder judgment on where to concentrate things. Hiberniantears (talk) 15:45, 10 December 2008 (UTC)
Thanks Seicer. I'm glad I'm not the only one with this problem. SMC (talk) 15:46, 10 December 2008 (UTC)
According security history on my computer, the similar attacks that I have received are from other URLs including, and even! Makes me think that this is either overly precautionary or perhaps a bug. LeaveSleaves talk 15:50, 10 December 2008 (UTC)

Heh, didn't think to look on the admin noticeboard for a technical issue. Sorry for the duplication. Anyway, I tend to agree that this is not something serious (I'm still editing here after all!), but something changed, and if it's on our end we should investigate. Powers T 16:11, 10 December 2008 (UTC)

Hum! dido. But I fear the change is on our end! Most likely Nortons, but I'dd like to know why they implemented this protocol? (If this is the case) The reason I think it's on our end is because I get the same error at the following Xerox website. website Let me see... I still need to reboot my computer and modem. I'll do that now. --CyclePat (talk) 16:15, 10 December 2008 (UTC)
Yep, I got the same intrusion attempt when attempting to respond to a post on the official EverQuest II forums. It seems to be edit boxes that trigger it for me, somehow. Powers T 16:39, 10 December 2008 (UTC)
  • Norton is incredibly paranoid. It's most likely a false positive. HalfShadow 17:43, 10 December 2008 (UTC)

Symantec info on this error. No idea why we are seeing a false positive here, I suspect they just borked their signatures for this type of attack. --TheDJ (talkcontribs) 16:48, 10 December 2008 (UTC)

This blog posting is a bit further in the research than us, and Wikipedia doesn't seem to be the only one affected (Google Maps). Seems like any "floating ajax" script can cause this signature to trigger. In wikipedia likely popups is the element to match the signature. --TheDJ (talkcontribs) 16:51, 10 December 2008 (UTC)
Norton Community Forum thread Symantec is aware of the issue and preparing a fix. --TheDJ (talkcontribs) 16:55, 10 December 2008 (UTC)
  • I still think this issue has something we should look into. The reason is because it appears to only be with English Wikipedia. Actually, it's funny. I just used the French Wikipedia, and Commons and, I don't get the error. Hence, I would like to know what is different between our English Wikipedia and the French Wikipedia which may be causing this problem? (Of course I should probably read the aformentioned blog posting. --CyclePat (talk) 17:04, 10 December 2008 (UTC)
  • Do you use popups, or any other scripts here that you don't use on fr.wikipedia? It seems the problem is related to script loading, so if you have no user scripts on fr.wikipedia, it likely won't trigger the problem. Mr.Z-man 17:58, 10 December 2008 (UTC)
Ah HA! Bingo! I logged in under my secondary account and no warning! Thank you. It must be something I installed a long time ago in my monobooks. --CyclePat2 (talk) 18:03, 10 December 2008 (UTC)
Oops! I think the monobooks may not be the issue. I just realized that the problem happens when I use Internet Explorer... I was using firefox (and still am without any notification error) when I logged in under my second account. (sorry! I forgot). IE give a bunch of errors. Perhaps yesterday's windows update touched IE? --CyclePat (talk) 18:09, 10 December 2008 (UTC)
I'm using Firefox, so it's not that simple. =) Powers T 18:14, 10 December 2008 (UTC)
Here I thought it might be Cumulative Security Update for Internet Explorer 7 in Windows Vista (KB958215) Installation date: ‎10/‎12/‎2008 3:08 AM. More information: (Part of the recent update) But I guess I'll have to wait and see. No matter the case the above regarding Firefox not causing me any problems still stands. But your statement regarding the use of firefox and having the problem has me agreeing... It's not that simple! Weird. --CyclePat (talk) 18:36, 10 December 2008 (UTC)
Probably best for those who don't *need* Symantec/Norton products particularly (i.e. non corporate standard operating environment) to go with their many competitors. As a former techie I saw its ability to completely finish off perfectly working machines. Orderinchaos 18:56, 10 December 2008 (UTC)
Please don't turn this into a Norton-bashing discussion. As a techie I've seen people who go with alternate products (usually AVG due to being egged on to install it by techie friends) have their machines completely ruined by all kinds of virii and trojan. It's enough for me to advocate installing it on every at-risk machine provided the *machine can run it*. It's good to see it's been fixed now, so I'll happily back out of this discussion. :) SMC (talk) 04:08, 11 December 2008 (UTC)
Actually, I think they did pretty well. According to this post on page 6 of the community forum of Nortons from a symantec employee, the problem has been resolved. Simply run live update. --CyclePat (talk) 19:08, 10 December 2008 (UTC)

(outdent) In fact Orla from Symantec stated in the afformention Community board: "Hello, I can confirm that the issues you've been seeing, where NIS erroneously triggers the detection "HTTP Acrobat PDF Suspicious File Download" on some legitimate websites, have been resolved. Corrected signatures have been created and are now available for download via LiveUpdate." They then appologized and indicated that they take these false reports seriously. --CyclePat (talk) 19:12, 10 December 2008 (UTC)

  • Confirmed that running LiveUpdate fixes the problem. At least it does for me. Nice work Symantec! Powers T 20:09, 10 December 2008 (UTC)
Nice work would have been to test their new signatures more througoutfully before putting them live. eBay, Google Maps are not 3rd class websites... -- lucasbfr talk 09:21, 11 December 2008 (UTC)

Tabbing when using cite button[edit]

When using the "easy cite" button (Button easy cite.png) on the edit screen and then picking a citation template, when one hits the tab key, it acts in a way that's less than intuitive. Currently, when one is in a field and then presses the tab key, rather than skip to the field immediately to the right (from the left field) or skip down and to the left field (from the right field), pressing tab moves the cursor to the main edit box. Thus to move between citation fields, one has to use the mouse - a colossal pain in the butt, in my opinion. Is it possible to change this to be more intuitive, to allow the tab key to move from one field to the next?

For what it's worth, I'm using Firefox 3 under Windows Vista. SchuminWeb (Talk) 00:12, 11 December 2008 (UTC)

Hmm, seems to behave as expected (and as you would wish it to) in Safari on WinXP. DuncanHill (talk) 00:14, 11 December 2008 (UTC)
I'm running FireFox 3 on WinXP. When I focus on the first box, tab takes me to the right box, then down left, right, etc. as expected. --—— Gadget850 (Ed) talk - 20:22, 11 December 2008 (UTC)
I just tried it at work, and it also exhibits this same behavior using Firefox 3 in Mac OS X Leopard. Works correctly on Safari on the same Mac, however. Perhaps this is a Firefox 3 problem? SchuminWeb (Talk) 22:12, 11 December 2008 (UTC)

"Undo" button acting weird[edit]

While attempting to revert some edits, I saw some very strange behavior from the "undo" button on the diff page; instead of just reversing the previous diff, it instead inserts a duplicate of the entire article in between the lead section and the first section header. Anyone know what's happening? rdfox 76 (talk) 00:15, 12 December 2008 (UTC)

Oh dear... confirmed, looking into it. —Ilmari Karonen (talk) 00:29, 12 December 2008 (UTC)
I've noticed this too. For me, it doesn't actually insert a duplicate of the entire article, it just looks like it does in the preview. -kotra (talk) 00:37, 12 December 2008 (UTC)
Yes, you're right. Still needs fixing, but it's not quite as serious as I thought it was. *phew*Ilmari Karonen (talk) 00:46, 12 December 2008 (UTC)
Is this related to the current malfunctioning of bots? Could be a MediaWiki change that broke a couple of things (I don't expect Cydebot and the AIV helperbots to develop similar problems at the same time). Kusma (talk) 00:40, 12 December 2008 (UTC)
There's nothing wrong with the bots, just MediaWiki's section-editing replacement function. Fixed in r44486 and deployed. --brion (talk) 00:57, 12 December 2008 (UTC)
Is it safe to turn my bots back on? --Carnildo (talk) 01:05, 12 December 2008 (UTC)
Pretty sure. :D --brion (talk) 01:19, 12 December 2008 (UTC)

Template dab help needed[edit]

The Template:Topic by country currently includes inks to Georgia, which is a dab page. It should link to Georgia (country). I haven't got a clue how to fix it, and did ask on it's talk page back in April, but no-one appears to have noticed. Anyone here able to do it? DuncanHill (talk) 02:07, 12 December 2008 (UTC)

I've more or less fixed it. For reasons beyond my comprehension (I'm pretty drunk), the links now point to Georgia(country) rather than Georgia (country), but at least they work. Algebraist 02:49, 12 December 2008 (UTC)
Many thanks. DuncanHill (talk) 03:06, 12 December 2008 (UTC)
I've fixed the redirect problem by replacing the spaces with underscores. Graham87 06:41, 12 December 2008 (UTC)

Add a css class to the fundraiser notices[edit]

Please add a descriptive css class like "fundraiserNotice" to the fundraiser notice (the one that's wasting a huge amount of screen real estate on everyone's screens on every article in the entire project for months at a time), so that we can block it with adblockers. I know I can block class siteNotice, but I don't want to block every site notice; just the fundraisers. —Preceding unsigned comment added by (talk) 20:36, 11 December 2008 (UTC)

It's #centralNotice --brion (talk) 22:03, 11 December 2008 (UTC)
Is #centralNotice used for anything except fundraisers? Algebraist 23:30, 11 December 2008 (UTC)

Not currently, but it's a generic framework for Wikimedia-wide central notices, so it could be in the future. — Werdna • talk 16:03, 12 December 2008 (UTC)

Please add a class that only applies to fundraisers. If centralNotice will be used for other things in the future, I obviously don't want to block it. —Preceding unsigned comment added by (talk) 16:12, 12 December 2008 (UTC)

MediaWiki version upgraded, "Image" namespace is now "File"[edit]

In case someone hasn't noticed it yet, we're now running MediaWiki r44485. The upgrade has brought about a bunch of changes, the most noticeable one being the renaming of the "Image" namespace to "File". Note that the old name "Image" will continue to work, so there's no need to go around changing existing links to use the new prefix. Some external tools and scripts may have broken due to the change: please report any breakage to their respective maintainers. —Ilmari Karonen (talk) 23:26, 11 December 2008 (UTC)

Should there be a list of things on a wiki locally that actually will break in user scripts, user styles, and parserfunctions? Here is what I can think of:
  • Parser functions like #switch and #ifeq comparing {{NAMESPACE}}. Most of these in en.wp and in system messages have been fixed aforethought by using {{ns:6}}
  • Javascript checking wgCanonicalNamespace (and to a lesser degree wgPageName).
  • Styles applied to individual pages in NS6, eg Not likely to be many of these.
Note that these are very peripheral, and chances are, you reading this? Yah, doesn't concern you. Move along citizen. *snide* --Splarka (rant) 07:25, 12 December 2008 (UTC)

"You are currently unable to edit pages on Wikipedia."[edit]

This is the second time this week I've been unable to edit Wikipedia because someone I don't even know decides to become a vandal. I'm not sure how they can possibly confuse my IP with there's but I'm guessing it's an ISP issue. Does anyone know of a possible way to avoid this in the future? Or maybe tell me why it's happening? I know about using the secure page but that's just a pain. Cheers. Rehevkor 03:42, 12 December 2008 (UTC)

Just to be sure what you are saying; You want to edit without logging in and you are blocked from editing. But if you log in, then you can edit. Is that correct? It is very possible that you are using a shared IP address and the only way around the problem is to log in. Dbiel (Talk) 03:51, 12 December 2008 (UTC)
I'm not sure what's happening. I'm editing logged on, and suddenly I get the warning saying I've been autoblocked. It seems because someone is vandalising at my IP but I'm home alone. I have to gather that it's an ISP problem, at least according to User talk: Anyone familiar with the issue? Rehevkor 03:54, 12 December 2008 (UTC)
Actually, just found this. Looks like my ISP is filtering Wikipedia, might explain why I'm through a different IP. Doesn't look like there's anything I can do. :/ Rehevkor 03:58, 12 December 2008 (UTC)
Thanks, the problem is autoblock see Wikipedia:Autoblock. I would bookmark that page and follow the instructions if it happens again. Dbiel (Talk) 04:08, 12 December 2008 (UTC)
(ec)You can see Internet Watch Foundation and Wikipedia and Wikipedia:Administrators' noticeboard/2008 IWF action‎ where you'll find details of the whole saga, and plenty of people to commiserate with. It's supposed to be pretty much resolved: the good folks at the talk page might offer some answers.Gwinva (talk) 04:11, 12 December 2008 (UTC)
For future reference, this probably could have been handled with an IP block exemption if it was just one isolated case. l'aquatique || talk 08:36, 12 December 2008 (UTC)


We need to be able to search for the first instance of a word, in all the history.

So, we want to find the first instance that say "Pilot" shows up in Cross-country skiing. We want to see this in history, so what we do is we search for history, but we want to see which version (or revision) this was typed, and thensaved, in. (talk) 08:12, 12 December 2008 (UTC)

WikiHistory can do this. It can be downloaded from the toolserver and needs the .NET Framework. --Oxymoron83 08:41, 12 December 2008 (UTC)
The tool labelled "Revision history search" in the page history can also do this. Here's a direct link to the tool. Graham87 11:08, 12 December 2008 (UTC)

Looking for help alphabetizing big list; is there a bot?[edit]

I have been working on List of skin-related conditions, and want to (1) alphabetize the diseases under each header, and (2) alphabetize the synonym names for diseases within each parenthetical, and wanted to know if there was a bot, or some other automated way, to do this? I would prefer not to do it manually. Thanks again for all your help! kilbad (talk) 04:34, 13 December 2008 (UTC)

Hello, please see list of active bots and/or post a request for a bot to do the task at WP: Bot requests. --ThaddeusB (talk) 05:37, 13 December 2008 (UTC)

Out-of-place div[edit]

User:Richard001 asked on the help desk about some of the links in his userpage not working. I checked it out with Firebug, and it's picking up the following code in the HTML source:

<div style="position: fixed; left: 40%; right: 10%; top: 10%; bottom: 10%;"/>

It's covering up exactly where two or three of his userboxes are, preventing you from clicking on the links in them (see the help desk post). The code does not show up in the source of the userpage given by Wikipedia [9] [10] or the HTML given by Special:ExpandTemplates. Where is this coming from (and why is it there)? Calvin 1998 (t·c) 04:46, 13 December 2008 (UTC)

Addendum: The problem code is in line 173 of the HTML source according to my browser. Calvin 1998 (t·c) 04:49, 13 December 2008 (UTC)
Screwy div now removed from template. {{Nihiltres|talk|log}} 07:26, 13 December 2008 (UTC)
Thanks. Calvin 1998 (t·c) 18:29, 13 December 2008 (UTC)

Need Help Controlling Image on an Article Page[edit]

I need help controlling the size and posistion of the image on this page: TikiTag Please let me know what I can do on my talk page. --Christopher Kraus (talk) 01:10, 14 December 2008 (UTC)

There, is that better? —Ilmari Karonen (talk) 01:18, 14 December 2008 (UTC)
Note, though, that the image appears to be non-free and potentially replaceable (since someone could take their own picture of the reader and tags). As such, it probably does not comply with our non-free content policy. What you might want to do is contact the company and ask if they'd be willing to license the image under a free license, such as {{cc-by-sa-3.0}}. —Preceding unsigned comment added by Ilmari Karonen (talkcontribs) 01:28, 14 December 2008 (UTC)

Age of database dump[edit]

The timestamp on the latest database dump says '2008-Oct-13'. The talk page of database download said the dumps were about weekly. What is happening? Lightmouse (talk) 15:53, 12 December 2008 (UTC)

The talk page is lying. They used to be weekly, but there's just so much stuff in the dumps now that they take weeks/months to run. — Werdna • talk 16:04, 12 December 2008 (UTC)
Maybe the devs should use Metamucil. ;) EVula // talk // // 18:16, 12 December 2008 (UTC)

Thanks. Are the downloads at predictable intervals? I am only interested in articles. Lightmouse (talk) 16:09, 12 December 2008 (UTC)

For the last few times, dumps arrived about every 1-2 months, however the part that exports the complete page history was usually cancelled. The current dump runs since 13 Oct as Lightmouse said. What I'm worried about is for how long it will run; the download page says "ETA 2009-06-16 10:53:02". Is that just an error, or is someone seriously considering to let the dump run until next June? --B. Wolterding (talk) 17:29, 12 December 2008 (UTC)
The dumps build off the the previous dumps, the problem is, the past several full page history+text dumps have all failed (I believe this is the "% prefetched" number, but I'm not positive), so a lot more work needs to be done. The problem is that the dumping script doesn't really recover from errors, so it takes 6 months to run, and one error can kill all the progress. Mr.Z-man 18:46, 12 December 2008 (UTC)
No, dumps don't build off of each other. And every full history dump for enwiki in the last ~2 years has failed. Dragons flight (talk) 18:48, 12 December 2008 (UTC)
Yes they do, see XmlDump.buildCommand in and the --prefetch option in /maintenance/dumpTextPass.php. Mr.Z-man 18:55, 12 December 2008 (UTC)
They use previous dumps as an input optimization, but are not reliant on them. The problem is simply that at our current scale, the old serial system for building the full-history dump takes too long and ends up encountering some problem eventually. It's to be replaced with a parallel process, which should hopefully be deployed within the next few weeks. --brion (talk) 00:05, 13 December 2008 (UTC)

I only need current page content so I can detect the string 'foobar' in an article today. I am sure there are lots of people running tasks just on page content regardless of history. A two month old dump is not an efficient source for detecting current content, particularly when the task needs repeating every few weeks. As the dump gets older, it becomes more effective to use the live server. Can we have a newer dump of current page without waiting for the full history dump? Lightmouse (talk)

A dump of any age will be of limited utility for working with *current* content, given the rate of editing. What sort of process are you doing, and is working with an offline dump actually the most efficient way to do it? --brion (talk) 00:07, 13 December 2008 (UTC)

I am looking for instances of broken date autoformatting where the year part of the date has been incorrectly redirected (e.g. [[April 11]], [[2005 in aviation|2005]]) thereby breaking autoformatting. On the basis of my current searches, I estimate that 1 article in every 1000 articles contains one or more of these errors. I have not used the database dump but I was under the impression that this would be an ideal first task. Lightmouse (talk) 00:45, 13 December 2008 (UTC)

I'll see if I can get the list for you using AWB. --NE2 06:27, 13 December 2008 (UTC)
Never mind; it was still recursing through Category:Years over two hours later. Once I got the list of "foo in year" articles, I was going to get a list of what links to them, then set it to skip if it contains a link of the sort, and skip otherwise. The tab of pages skipped can then be sorted by why it was skipped. --NE2 08:43, 13 December 2008 (UTC)

Yes, I also tried that with AWB and failed. Thanks anyway. Lightmouse (talk) 11:20, 13 December 2008 (UTC)

This is another case where storing the "pipe" text in the `pagelinks` table would be helpful. Then it could be made visible in Special:Whatlinkshere/2005_in_aviation or whatever. — CharlotteWebb 00:39, 15 December 2008 (UTC)

Table help![edit]

Can someone help me with the table at Lancashire Council election, 2009? I would like the first column to be split in two - one colour box for the party, one column for the party name/link. I keep adding "!" or "|" but this just knocks the table out of synch. Any ideas? I half recall "col=2" or something...? Thanks doktorb wordsdeeds 08:25, 13 December 2008 (UTC)

Problem fixed. It was the | above the ending |} that produced the odd-looking table. The "pipe" implies that there's another column to be added, and that's why an empty strip appeared on the right side. I believe you were not thinking of "merging cells" or the like right?
Sorry to post something unrelated here, but I urge those really technical person to take a look at my issue here. Hytar (talk) 06:00, 14 December 2008 (UTC)
Looking at the code you provided, assuming it's not part of a larger template, you could remove {{for loop}} and instead just put the || between each table cell, which would save a few characters and make the template easier to understand. I noticed the template {{Format price}} doesn't exist. Assuming you mean {{formatprice}}, one option to save characters would be to make another template which adds two numbers together and formats them, which you can then use instead of using #expr: multiple times in your code. This would only be necessary however, if you're making a lot of similar templates. Tra (Talk) 11:25, 14 December 2008 (UTC)
First, thanks for responding! The thing is, I wanted to simplify the table-making process by adopting a template, but that's just for one article that I'm concerned. I wanted to design it so that with a few inputs, a table can be generated "Excel-style".
The article concerned is the pricing section of the List of Astro channels. Hytar (talk) 22:53, 14 December 2008 (UTC)

Thanks for the help :) doktorb wordsdeeds 19:10, 14 December 2008 (UTC)

Signature & Date Stamp[edit]

If I'm logged out I can sign my user name, but if I'm logged in I can't. None of the other buttons on the bar above work either. I will now log out and sign, but follow the link to see my user page.-- (talk) 10:29, 13 December 2008 (UTC)

Check your preferences, at a guess it looks like you either have screwed up your signature, or have some conflicting gadgets. --Splarka (rant) 09:14, 14 December 2008 (UTC)

Thanks, Splarka. It now works, even though it stopped working when I hadn't changed anything. I checked the references and unclicked/clicked a few things, so maybe it was just one of those things that happen, who knows? Thanks again.--andreasegde (talk) 12:11, 14 December 2008 (UTC)

CSS Code for iPhone Compatibility[edit]

Moved to MediaWiki talk:Common.css Happymelon 20:08, 14 December 2008 (UTC)

Source tag syntax highlighting[edit]

When writing programming source code in articles, it seems pretty common to use the source tag for automatic syntax highlighting. However, the result is often not optimal. Consider the following example in the C++ programming language:

// ok, just a comment

#include <iostream> // not ok, this comment has a different color 

int main() // ok, 'int' is a keyword
  if(not false) // maybe ok, 'if' and 'false' are keywords, 'not' is a reserved word
    std::cout << "this is a literal"; // not ok, 'cout' is not a keyword

I just wonder where these syntax highlighting rules are defined, and whether it's possible to have them changed.Ufretin (talk) 03:02, 12 December 2008 (UTC)

This functionality is provided by mw:Extension:SyntaxHighlight GeSHi. The source code is here. Algebraist 03:05, 12 December 2008 (UTC)
We can define custom Geshi formatting in MediaWiki:Geshi.css. I think a set of rules to make all languages render in broadly similar colours would be a step forward. Happymelon 11:06, 14 December 2008 (UTC)
The issue is not the coloring used, but questioning of details of the highlighting rules -- appropriateness of entries on the keyword list and whether comments should be marked separately within preprocessor directives. --brion (talk) 15:44, 15 December 2008 (UTC)

Internet images not showing up[edit]

What might be the cause of an internet page not showing the images when the page is loaded in a web browser? —Preceding unsigned comment added by Freshb28 (talkcontribs) 15:57, 13 December 2008 (UTC)

A slow connection somewhere in the pipeline? If the connection to the image server was slowed down by something (not necessarily on your end), that could cause the page to load but the images not to. EVula // talk // // 19:07, 14 December 2008 (UTC)
In Firefox it's also easy to accidentally click the "block images from this server" in the right-click menu. :) --brion (talk) 15:40, 15 December 2008 (UTC)

Section edit links?[edit]

thumb|right|screen cap of section edit links When did the section edit links get moved to the left of section headings? Thanks. – ukexpat (talk) 05:16, 14 December 2008 (UTC)

I still see them where they've always been (on the right). §hep¡Talk to me! 11:21, 14 December 2008 (UTC)
Same here. EVula // talk // // 19:08, 14 December 2008 (UTC)
I have uploaded a screen cap, thumbnailed at right.  – ukexpat (talk) 20:49, 14 December 2008 (UTC)
What operating system (Windows XP? Mac OS 10.4?) and browser are you using? Anything unusual in your preferences - particularly the use of a complicated gadget like wikEd, Twinkle, or Friendly? -- John Broughton (♫♫) 21:18, 14 December 2008 (UTC)
That is firefox running Windows XP. Ukexpat- first try purging your cache and reloading a page. Secondly, if we aren't able to figure it out, here's a javascript snipped that, added to your monobook, should move the edit links to the other side of section titles.
Hope this helps- l'aquatique || talk 21:30, 14 December 2008 (UTC)
Thanks for the code, that put them back where they were at the right hand side of the window. Clearing the cache fixed it without needing your code, but thanks for suggesting it. I still do not understand why this only started happening yesterday - was there a mediawiki code change that only affects Firefox that was quickly reverted?  – ukexpat (talk) 22:45, 14 December 2008 (UTC)
I don't know, maybe although you'd think someone else would have noticed. Maybe it was a bug in your browser. Glad to hear it's fixed. l'aquatique || talk 06:46, 15 December 2008 (UTC)

Has anyone ever come up with statistics for article referencing?[edit]

Not sure if this is the right place to ask, but has anyone ever created a bot or whatever that counts or estimates the average number of references per article. Thanks. - Peregrine Fisher (talk) (contribs) 01:25, 15 December 2008 (UTC)

Although WP:BOTREQ is primarily for requests for a bot to actually do something, I don't think that this question would be out of place there. -- John Broughton (♫♫) 21:15, 15 December 2008 (UTC)
Thanks for the suggestion. I'll try that. - Peregrine Fisher (talk) (contribs) 21:29, 15 December 2008 (UTC)

Resource (Windows) article is broken or something...[edit]

Every time i try to view this article, a "would you like to download this unknown file" prompt appears. I've cleared my cache i-dont-know how many times and even tried reinstalling internet explored and firefox. It doesn't matter. Can someone fix this? I can't even get access to the page to try.  Buffered Input Output 14:23, 15 December 2008 (UTC)

Actually, it only works sometimes. it just worked for me now.  Buffered Input Output 14:23, 15 December 2008 (UTC)

Seems to work in Firefox/Ubuntu. What operating system, browser, and javascript addons do you have? l'aquatique || talk 20:07, 15 December 2008 (UTC)

Collapsed Table of Contents[edit]

All of a sudden the Tables of Contents in articles and talk pages are collapsed. I'm not sure I like it. I don't see it as an option in my preferences. Is there a discussion of this change? ·:· Will Beback ·:· 21:44, 15 December 2008 (UTC)

I don't see that; the ToCs are collapsible, and have been for at least 18 months, but I don't see them collapsed by default, which I assume is what you mean. Can you show an example/screenshot? Happymelon 22:01, 15 December 2008 (UTC)
If the TOC is collapsed, it will remain that way on all articles until you expand it, in which case it should remain expanded. EVula // talk // // 22:02, 15 December 2008 (UTC)
Hmm... maybe it's the cosmic rays. In any case, it sounds like it's not affecting others. Thanks for the responses. ·:· Will Beback ·:· 23:18, 15 December 2008 (UTC)

File: vs. Image:[edit]

I just noticed that all our images begin with "File:" now, rather than "Image:". Just curious (1) when this change was made, and (2) what effect does it have on me, who places and moves images sometimes but never does anything with non-image files? Nyttend (talk) 01:43, 12 December 2008 (UTC)

It should not really have any effect as the File: and Image: prefixes are synonymous except that "File:...." is displayed when you go to an image page. See bugzilla:44 and rev:43639 for details. Icewedge (talk) 01:57, 12 December 2008 (UTC)
The change was made about three hours ago, and it shouldn't affect you. --Carnildo (talk) 02:03, 12 December 2008 (UTC)

Why was this change made? Will it affect the way images are displayed or are we to place them in articles as File: not Image: from now on? The Bald One White cat 13:34, 12 December 2008 (UTC)

Because it was thought to be silly to use the prefix Image: for files which are not images. No, it will not affect you. Algebraist 15:51, 12 December 2008 (UTC)
This change has been planned for several years, but just got held back until people had time to track down and work around the internal compatibility issues that would be created by changing the namespace.
We've switched it from "Image" to "File" because it's pretty weird-looking for *non*-image files such as video and audio clips, PDFs, and office documents to be labeled as "Image"s, whereas there's nothing weird about labeling all files as "File"s.
"Image:" and "File:" now do exactly the same thing; there's no change in how they're displayed. You can use "File:" or "Image:" depending on your preference. --brion (talk) 15:52, 12 December 2008 (UTC)
Woohoo! I knew that we would have to switch eventually, but I didn't expect it so soon. And it was so smooth I almost didn't notice. Hey, it seemed to me that I had to search all over to find this explanation - shouldn't we all be celebrating? I guess the only downside is some housecleaning will be required - for example, meta:Help:Image page. Anyway, good work. –Sarregouset (talk) 18:12, 12 December 2008 (UTC)
Huh, I had figured it had something to do with the IWF situation earlier. —/Mendaliv//Δ's/ 19:12, 12 December 2008 (UTC)

Question: How does this fit in with the Media namespace? Media has seemed to be a synonym for Image for a long time, and now all three of them work (wgNamespaceNumber: 6, wgCanonicalNamespace: File). The weird thing is that if you access a page with the Image prefix it issues an HTTP 301 redirect to the corresponding File page but if you use Media, it doesn't. E.g., Image:Tile_Hill_train_550.jpg, Media:Tile_Hill_train_550.jpg. This sort of thing is critical for stats tracking, which needs to distinguish between actual views and mere redirects. Why/when/how was Media introduced anyway, and why is it different? • Anakin (talk) 15:22, 16 December 2008 (UTC)

I sort of like "Media:" better, actually. It's more descriptive than "File:".--Pharos (talk) 15:33, 16 December 2008 (UTC)
Odd but if "Media:" was going to http-redirect to anything, shouldn't it be the deep link for the image itself <> (to mimic the behavior of the wiki-link "Media:Tile Hill train 550.jpg")? Or maybe we don't want to make it that much simpler for other sites to leach bandwidth (but they already do it anyway, shrug). — CharlotteWebb 15:36, 16 December 2008 (UTC)
You can deep link with Special:Filepath, both as a wikilink Special:Filepath/Tile Hill train 550.jpg and as an external link:
"Media" probably doesn't 301 because it is a real namespace (although when you are on a Media: namespace page you're in namespace 6, and not -2), instead of an alias like "Image" is now. It was first added 5 years ago. But I agree, currently there are way too many ways to link to an image or file. --Splarka (rant) 16:49, 16 December 2008 (UTC)
Nope that's just a bug, it should redirect to the canonical form like everything else, it just needs to be special-cased due to its being special. bugzilla:16677 --brion (talk) 19:36, 16 December 2008 (UTC)
So there's no chance the IWF will figure out all of them. 718smiley.svgCharlotteWebb 17:14, 16 December 2008 (UTC)
They don't care anymore. --Splarka (rant) 19:16, 16 December 2008 (UTC)
That is an amazingly confusing list, Splarka. I don't really get why was "Image" renamed to the completely new namespace "File" when "Media" was already in use. The complexity of having 3 namespaces which almost but don't quite mean the same thing doesn't seem to be at all justified by the desire for more accurate titling of files. • Anakin (talk) 17:54, 16 December 2008 (UTC)
Media: is used to link to a specific piece of media, while Image: (and now File:) are used to embed the object, or link to the description page. They do completely different things. Media: when used as a wikilink actually creates an inline link to the file. When used in something like {{#fullurl:media:Wiki.png}} it acts like a wiki #REDIRECT causing the contents of the actual page to be transcluded, but this isn't the primary use of Media: (contrarily, Image: now http redirects, as it is an alias). Another useage difference is with #ifexist, for example: {{#ifexist:File:Foo.ext}} will tell you if such a description page exists in the image namespace, but not if there is a file associated with it (you can create arbitrary NS-6 pages). {{#ifexist:Media:Foo.ext}} will tell you if there is a media file with that name (which can exist from commons, for example, independent of the description page). Using both can tell you the status of the image: local, commons, page without image, or missing. So you see, we need both. --Splarka (rant) 19:16, 16 December 2008 (UTC)
Thanks Splarka, that makes sense now. {{#ifexist:Media}} for detecting Commons images sounds useful too. • Anakin (talk) 00:36, 17 December 2008 (UTC)
Is there any other case where an internal link [[Foo:Bar]] and an external link {{fullurl:Foo:Bar}} behave differently (assuming the page exists rather than inviting you to create it)? Something doesn't seem right… — CharlotteWebb 18:12, 16 December 2008 (UTC)
Interwiki links are treated a bit odd. If they are wikilinks or {{fullurl}} links, the first two can be non-local (only local interwikis get the off-site http bounce). For example, compare mw:foo and {{fullurl:mw:foo}} to This is because the first depth can be parsed and actually create an offsite URL. The 'bounce' http redirect is especially needed for stacked interwikis, as long as they are local. For example, two locals can bounce: wikt:fr:foo or the reverse fr:wikt:foo. Compare that to m:mw:foo vs mw:m:foo (Meta is local but MediaWiki is not). Confused yet? --Splarka (rant) 19:16, 16 December 2008 (UTC)
Image:/File:, Media:, Category:, and interlanguage links all do special things when linked directly inline. --brion (talk) 19:36, 16 December 2008 (UTC)
Err yeah, those are formatted differently (in the sidebar for interwikis, etc.) but still link or redirect to the same url. That's what I meant. — CharlotteWebb 19:58, 16 December 2008 (UTC)

New Magic Word[edit]

I was going to post this at one of the various Help:Magic Words articles (on Wikipedia, Meta, and MediaWiki), but it seemed that no one had been to those pages in ages (1 year+), so I didn't think my question would be answered there. Magic Words ({{CURRENTTIME}, {{PAGENAME}}, etc.) trigger a function in the Wikimedia software to do something. I would like to propose a new Magic word, {{PAGEWIDTH}}. This function would use the javascript parameter window.innerWidth (on IE, document.documentElement.clientWidth or document.getElementsByTagName('body')[0].clientWidth) to find the width of the window. This would be useful in many templates to determine the size of things such as images, tables, etc. If someone wanted to set an image to display at full width of a column in a page (say a 50% column), the user could use this magic word and a combination of expr functions to find out exactly how many pixels this image should be displayed at. The javascript code is shown below:

<script type="text/javascript">

// initially set to 600 (lowest resolution we need to worry about) in case none of it works

 var windowwidth = 600;
// FF/Opera/IE7/Safari use window.innerWidth
 if (typeof window.innerWidth != 'undefined')
      windowwidth = window.innerWidth,
// IE6 with a valid doctype as the first line in the document uses the following

 else if (typeof document.documentElement != 'undefined'
     && typeof document.documentElement.clientWidth !=
     'undefined' && document.documentElement.clientWidth != 0)
       windowwidth = document.documentElement.clientWidth,
// older versions of IE use this
       windowwidth = document.getElementsByTagName('body')[0].clientWidth,

The window width is now in the variable windowwidth and is outputted. --Dudemanfellabra (talk) 23:51, 14 December 2008 (UTC)

I'm fairly sure this can't be done. Magic words have their values substituted when the article text is parsed by MediaWiki, while you don't know the value of the window width until someone tries to view the page in a web browser. --Carnildo (talk) 00:24, 15 December 2008 (UTC)
I'm not quite sure I understand what you're talking about. MediaWiki parses the article text when and only when someone tries to view the page in a browser, right? I mean CURRENTTIME refreshes with each reload of the page; so would PAGEWIDTH. The only downside is that if someone resized the browser after it had been parsed, the magic word wouldn't be dynamic, so that image or whatever it was applied to would stay the same width and not scale with the page.. but all they'd have to do to fix the problem is refresh the page after they had resized their window. --Dudemanfellabra (talk) 00:31, 15 December 2008 (UTC)
No, the magic words are expanded when the page is parsed and served to the user (if the parsed page is not cached). Javascript, on the other hand, executes on the client after the text is fetched from the server. So there is really no (feasible) way to implement this as a magic word at the MediaWiki parser level. However, as you've pointed out that a solution (if indeed this really is a problem...) exists in javascript, you could amend MediaWiki:Common.js to include a function such as the one you provide to substitute a variable in the parsed html served to the client with the page width. Naturally, however, you'd need to get consensus for such a change, ensure compatibility across browsers, etc. AmiDaniel (talk) 01:37, 15 December 2008 (UTC)
It is possible to convert the magic word to a SCRIPT element that document.writes the calculated value client-side; it would not work without client-side script support though. What is it needed for? --Yecril (talk) 18:23, 15 December 2008 (UTC)

Not technically possible. — Werdna • talk 03:10, 15 December 2008 (UTC)

The <img> tag supports percentage-based widths, given that images are really the only use for this, it would make a lot more sense to change the image parsing code to accept a % width in addition to pixels, rather than a hacky JS-based magic word, that if used in an #expr would just break horribly for people with JS disabled. Mr.Z-man 17:17, 16 December 2008 (UTC)

Never mind, ignore what I just said. Most image scaling is done server-side, so setting width% wouldn't work. But the proposed "magic word" is still far too much of a hack. Mr.Z-man 18:35, 16 December 2008 (UTC)

How many deletion edits[edit]

  1. Is there anyway to quantify how many edits in, lets say in a day, are in AfD/RfD/MfD and other related articles related to deletion?
  2. Is there anyway to quantify how many edits were in those articles which are up for deletion?
  3. Is there anyway to calculate the average amount of time it takes an editor to make one edit?

Thank you. travb (talk) 06:06, 15 December 2008 (UTC)

I think MfD/RfD would be relatively easy, since they're split by day, you could just look at that subpage's history and count how many edits were made in a given day. AfD uses a separate subpage for each article, so getting a daily count there would be trickier but by no means impossible. The article bit would be trickier, especially if you want to know the number of edits while up for deletion as opposed to the number of edits the article has gotten in general. Is there something in particular you believe this data would be useful for? Seraphimblade Talk to me 06:20, 15 December 2008 (UTC)
A slight correction: MfD also uses a separate subpage for each entry, just like AfD. Graham87 07:52, 15 December 2008 (UTC)
Thank you, I didn't know it would be that easy, I thought it would involve looking at the wikipedia back up logs that acedemics always use for studies.
I was hoping to count all edits to pages related to deletion. For a given week, for example.
How can I calculate the average amount of time it takes an editor to make one edit? I have some vague ideas...travb (talk) 14:30, 15 December 2008 (UTC)
"Average" is probably not the correct measure for this -- I expect that the time taken to make an edit will follow a decaying exponential curve, ranging from huge numbers of ten-second spelling fixes to a small number of multi-hour edits creating entire articles. --Carnildo (talk) 22:57, 15 December 2008 (UTC)
Median or Mode then? Which one? travb (talk) 04:58, 16 December 2008 (UTC)
Depends on what you're doing with the number. --Carnildo (talk) 22:22, 16 December 2008 (UTC)

Polish grammar forms[edit]

MediaWiki:aboutsite/pl reads O Wikipedia instead of O Wikipedii.

This should be fixed by adding

$wgGrammarForms['pl']['MS.lp']['Wikipedia'] = 'Wikipedii';

to configuration settings.

--Yecril (talk) 18:09, 15 December 2008 (UTC)

Please use --TheDJ (talkcontribs) 00:38, 16 December 2008 (UTC)

I disagree. It is not a MediaWiki issue because Wikipedia is not a MediaWiki word. It would be insane to expect WikiMedia to support the declension of every possible site name in every possible language. --Yecril (talk) 09:05, 16 December 2008 (UTC)

The MediaWiki Bugzilla is used to track issues with Wikipedia's infrastructure. Don't let the name confuse you. Chris Cunningham (not at work) - talk 10:08, 16 December 2008 (UTC)

Bugzilla:16669 --Yecril (talk) 11:20, 16 December 2008 (UTC)

404 link on a Special: page[edit]

In Special:LinkSearch, the Crosswiki linksearch link throws a 404 error. --Enric Naval (talk) 19:57, 15 December 2008 (UTC)

That is run by User:Nixeagle so I would suggest asking him. MBisanz talk 17:21, 16 December 2008 (UTC)

Watchlist cruft[edit]

Would it be possible to create a gadget that permanently disables the watchlist noticecruft? For various technical and security reasons I am sometimes not using cookies or CSS when looking at my watchlist, and in particular, on a mobile device it is annoying to have to wade through the list of everybody's pet project just to get to my watchlist. The user interface should not be used to intrusively propagate notices, but until people get that point, can we please provide a work around. That will be much better than endless debates and edit warring over the notice-spam. Jehochman Talk 19:58, 15 December 2008 (UTC)

Just add:
#watchlist-message { display:none; }
to your monobook.css (or the .css page for whatever skin you use). Mr.Z-man 20:20, 15 December 2008 (UTC)
That does not work on mobile devices that don't support CSS, or don't load CSS for the sake of appearance or efficiency. I want my user interface to be clean. I don't need a multiple screen bulletin board appended to the top of my watchlist. Jehochman Talk 22:07, 15 December 2008 (UTC)
If you want to hide it, you need CSS, simple as that. Seriously if you don't use CSS at all this site looks diabolically awful (try it, it's ugly) anyway; a cluttered watchlist is the least of your worries. There is no way to hide this content without using either CSS, Javascript or cookies. Happymelon 22:59, 15 December 2008 (UTC)
Not loading CSS for the sake of appearance? That sounds rather illogical. But as Happy-melon said, there's no way to hide it without CSS or JS (short of redoing the whole watchlist message thing to load the messages using AJAX, which would just be pointlessly complex). Mr.Z-man 02:35, 16 December 2008 (UTC)

If your device loads javascript but not css, you could try brute force:

document.getElementById("watchlist-message").setAttribute("style", "display:none;");

However I can't imagine why this would be the case. — CharlotteWebb 05:14, 16 December 2008 (UTC)

Hmm. It seems that my mobile device interprets some, but not all CSS definitions. Does the site have handheld media CSS definitions? Perhaps those are the ones being loaded. Jehochman Talk 22:03, 16 December 2008 (UTC)
Yes, MediaWiki:Handheld.css. Happymelon 22:13, 16 December 2008 (UTC)

I need help converting image formats[edit]

For the life of me, I couldn't find the proper forum for asking this question, so I guess here has to do.

I have shapefiles of Oregon Legislative Assembly districts, which can be found at the very bottom of this page. I'm very certain that these files can be used to make vector maps, but I simply cannot find a method available to me in order to do that. I Google endlessly, and I find several shp2svg converters, but they're either not free for me to use, or impossible for me to figure out how to use.

So, I'd very much appreciate it if someone could help me figure out how to accomplish this task, or even volunteer to do it themselves. If no one here knows how to help, please tell me where I can go to find people who do. I'm really desperate! Äþelwulf Talk to me. 01:27, 16 December 2008 (UTC)

I can't answer your question, but I can suggest two people who might be able to: Jmchuff (talk · contribs) (a fellow Oregon type), who's done some excellent maps based on Tri-met-supplied data, and Kmusser (talk · contribs), a professional cartographer who's made some excellent watershed maps and, probably, a whole lot of other kinds of maps too. As for free tools, I hear that Grass is pretty capable, but not the easiest to figure out. Might be worth hunting around for a Grass user community, maybe outside Wikipedia? -Pete (talk) 02:38, 16 December 2008 (UTC)
Thanks Pete. I'll consult them now. Äþelwulf Talk to me. 03:48, 16 December 2008 (UTC)
Copyright notice to anyone who might be interested in helping: It's my belief that the shapefiles I linked to are not copyrighted, although legislative district maps do exist that were published by the state government and are copyrighted. In this discussion at WP:MCQ, I reached the conclusion that the data about these legislative districts is not copyrightable, so the shapefiles are free, though the derivative works could be under copyright if they look too similar to the state government's maps. Links to those maps can be found in the aforementioned discussion at WP:MCQ. Äþelwulf Talk to me. 04:20, 16 December 2008 (UTC)
Also, to be clear on what I'm looking for, I'm interested in simple blank maps showing the district boundaries, which can be colored for various purposes, similar to Oregon's county maps like this one. Äþelwulf Talk to me. 04:44, 16 December 2008 (UTC)
You also might try the graphics lab, WP:LAB. l'aquatique || talk 04:46, 16 December 2008 (UTC)
Files converted (available on my talk page), I did it using GIS (namely ArcGIS), I think you do need a GIS program to do this conversion (especially since in this case I also had to do a map projection change) - there are freeware GIS out there, but I don't know how user friendly they are.Kmusser (talk) 20:10, 16 December 2008 (UTC)

Bot for building a project watchlist[edit]

Is there currently a bot that builds a project watchlist? This would be a generated list of all the talk pages that our project templates are placed on, as well as the article or other types of pages the talk pages are attached to. WatchlistBot used to do it, and then JoshurBot did it, but I don't know which bot does it now (if there is such a bot). Stevie is the man! TalkWork 05:59, 16 December 2008 (UTC)

I have not used it, but check out Igor. --—— Gadget850 (Ed) talk - 21:36, 16 December 2008 (UTC)

Recent changes on several pages tied together[edit]

Is there a way to look up recent changes to links listed on several pages, combined together? Stevie is the man! TalkWork 18:13, 16 December 2008 (UTC)

Transclude them all on one page, like so. The only problem is that you won't be able to keep it in that state if there are any mainspace categories. --NE2 18:20, 16 December 2008 (UTC)
Thanks. I don't believe mainspace categories are involved in my case. Stevie is the man! TalkWork 18:35, 16 December 2008 (UTC)
Maybe we need a __REMOVECATS__ tag. This would also be useful if you are working on a sandbox copy of an article or explaining how to use a stub template, etc. (to clear all categories appearing before the tag), shrug. — CharlotteWebb 18:43, 16 December 2008 (UTC)

Talkpage control[edit]

Ok, I just got ended up on one of grawps unblock pages, again, A Nu Wikipedian. A Wiki Nupedian, which is why I'm here. If've you've ever come across one of these pages, you'll know how much of a bitch they are, they take ages to load, block up the browser, then you have to mess about typing the address to get to the history so you can delete it. If you haven't seen one, its a huge image, along with other crap, all done using tables, and coloured cells, all adding up to about 2 million bytes in size. Now if you actually look at the code for these pages, it uses hundreds of these, <TD BGCOLOR=#6b4b2c>....</TD>.. Is there any possibility of limiting the number of times these can be used on a page somehow, as far as I know they are barely used anywhere else. Failing that, can anyone think of any other technical way of preventing the creation of these pages, if we can stop these grawp loses another "weapon"--Jac16888 (talk) 21:34, 16 December 2008 (UTC)

You can easily go to the page history by using {{page history link}}. -- IRP 21:47, 16 December 2008 (UTC)
I believe mw:Extension:AbuseFilter is intended to deal with this sort of thing. Of course, it hasn't been activated yet, so that doesn't help too much. Anomie 21:55, 16 December 2008 (UTC)
You can get to the page history by pressing the access key with h (alt-h in Internet Explorer, alt-shift-h in Firefox, probably a different key combo for the Mac). Graham87 23:57, 16 December 2008 (UTC)

History of good or featured articles[edit]

So many articles have page after page of minor edits, but is there, or could there be, a tool to allow one to see the expert or experts who had the most contributions to an article that really meets, or comes close to meeting, the standard for a good or featured article?Vchimpanzee · talk · contributions · 21:59, 16 December 2008 (UTC)

You can get a list of who has made the most contributions to an article with the "revision history statistics" link in each page history. There is no flag in the software that signifies a user's expertise in any topic, so you'll have to decide for yourself who the subject experts are. Graham87 00:00, 17 December 2008 (UTC)

Placement of deletion drop down bases is broken[edit]

The drop down link menu for deletion bases when performing deletions is no longer automatically choosing the bases associated with the deletion tag, i.e., when you attempt to delete an article tagged with db-bio, the reason for deletion field no longer automatically places, "A7 Bio: No indication that the article may meet guidelines for inclusion". Anyone know what's causing this?--Fuhghettaboutit (talk) 13:38, 9 December 2008 (UTC)

Most likely cause is something you've changed in User:Fuhghettaboutit/monobook.js recently, breaking MediaWiki:Sysop.js. Check your error console or try another skin temporarily (like action=delete&useskin=myskin ). --Splarka (rant) 08:21, 10 December 2008 (UTC)
Thanks for the explanation. The issue simply subsided after a day.--Fuhghettaboutit (talk) 17:49, 13 December 2008 (UTC)
I've had this problem today myself, and I haven't touched any of my monobook files recently. JPG-GR (talk) 06:47, 17 December 2008 (UTC)

ISO language template[edit]

Does anyone know of a template or magic word that converts ISO 639 langauge codes into the English equivalent of the language (e.g. fr to French, de to German) etc. The magic word {{#language:}} is the best I've found so far, but it converts to the local equivalent of the language (fr to français for example). —  Tivedshambo  (t/c) (logged on as Pek) 14:37, 16 December 2008 (UTC)

meta:Template:N en and meta:Template:Language do this. You could copy one of them across to en.pedia if there's nothing here already. Algebraist 15:44, 16 December 2008 (UTC)
One fix would be to make the parser-function two-dimensional. It would be quite useful to add an optional second parameter for localizing the name (use native name by default):
  • {{#language:en}} = English
  • {{#language:fr}} = Français
  • {{#language:en|fr}} = Anglais
  • {{#language:fr|en}} = French
Then we wouldn't need a separate lang-foo template for every language. — CharlotteWebb 15:55, 16 December 2008 (UTC)
There is an extension written by betawiki:User:Nike integrating CLDR (part of which is a project to translate every ISO 639 language name to every other language name apparently, which is far from complete) into MediaWiki. The extension can be seen in /extensions/cldr/. --Splarka (rant) 17:17, 16 December 2008 (UTC)
At OmegaWiki you will find collections for the ISO-639-1 and the ISO-639-3. Thanks, GerardM (talk) 17:26, 16 December 2008 (UTC)
Well, I guess it would default to the native name if there is no mapped value or for any unrecognized language code. First step might be to scrape the interwiki links from enwiki's article about each language (filtering out each language's word for "language"—we might want a function to give us that info too Smiley.svg). That still wouldn't be complete and there may be a few errors but it would be a good start and could be merged with anything done manually. Despite what Andre says [11] letting the projects build the software isn't inherently bad. — CharlotteWebb 17:39, 16 December 2008 (UTC)
There are 34,000 different intersections in a two dimensional function. Useful though it is, I'm not surprised no one's tried to implement that yet. Happymelon 17:41, 16 December 2008 (UTC)

Might be a handy user-prefs option for the sidebar too. — CharlotteWebb 17:45, 16 December 2008 (UTC)

I've created {{N en}}, based on meta:Template:N en. Thanks. —  Tivedshambo  (t/c) 16:53, 17 December 2008 (UTC)
I'm going to repurpose {{Language}} as an attempt to create that 2-D function discussed above. Happymelon 18:06, 17 December 2008 (UTC)


Can anyone explain to me how this works? I'm working on some javascript in a file on my hard drive that would use this function, but I can't find it defined anywhere. I tried adding <script type="text/javascript" src=""> (which appears in the source code of Wikipedia's main page) to my local file, but that doesn't change anything. I know the javascript works because when I add it as an onload function, everything works perfectly (i.e. <body onload="func()">). The problem only occurs when I try to take it out of the body tag and add in the addOnloadHook() function. My only guess is that the function is defined somewhere else, and I just can't find it. If someone could post the code for this function or simple give me the link to the js file in which it is defined, that would be awesome. Thanks! --Dudemanfellabra (talk) 22:29, 16 December 2008 (UTC)

Have you seen Wikipedia:WikiProject User scripts/Tutorial? -- Rick Block (talk) 01:07, 17 December 2008 (UTC)
(edit conflict) The addOnloadHook function is defined in wikibits.js. It takes one argument, a function. All functions given as arguments to addOnloadHook are called at the end of the <body> tag by another function from wikibits.js, runOnloadHook. This is used for almost all scripts that need to access the DOM, since many parts of these scripts need to be delayed until the page is done loading. Here's the code of the function:
function addOnloadHook(hookFunct) {
    // Allows add-on scripts to add onload functions
    if(!doneOnloadHook) {
        onloadFuncts[onloadFuncts.length] = hookFunct;
    } else {
        hookFunct(); // bug in MSIE script loading
Pyrospirit (talk · contribs) 01:18, 17 December 2008 (UTC)
Hmm.. well I have the code in my monobook.js now, so it works, but when I had it in my external editor, it wouldn't. I had wikibits.js linked to in the style part of the page, and I used the addOnloadHook() function correctly, so I'm not sure why it wouldn't work. You can look at the code on that page, and the only thing I had different was I added
<script type="text/javascript" src="">
my code
The code would would only work if I took out the addLoadEvent() function and added an onload="ScreenInfo()" to the body tag. Any ideas why? --Dudemanfellabra (talk) 02:54, 17 December 2008 (UTC)

As stated above, you need the script at the end of the mediawiki <body> in yours:

<script type="text/javascript">if (window.runOnloadHook) runOnloadHook();</script>

Also, you may need to define some global variables above the wikibits.js load to prevent the code in it immediately executed from causing various errors. IIRC these are stylepath, wgContentLanguage and wgBreakFrames. Here is an example. --Splarka (rant) 08:09, 17 December 2008 (UTC)

Ah, I see now. I guess I just skipped right over that part in the previous reply. Thanks, guys! --Dudemanfellabra (talk) 08:46, 17 December 2008 (UTC)

Loading revision history comparisons[edit]

I have never had problems viewing any parts of Wikipedia before (except perhaps some Unicode), but I recently switched to the green-and-black skin for Wikipedia, and since then I have been unable to view pages from the revision history of any article, on IE or Firefox, even after un-checking the gadget (and reverting to the default skin). Whenever I click on "Compare Selected Versions," "cur," or "last," it brings up a prompt asking me if I want to save the .PHP file or open it using a program from a list. I have tried several circuitous routes to opening it, but it seems my computer simply cannot recognize or handle the files. When I tell it to let Firefox or IE open it, it simply opens a new, blank tab giving the same prompt (and if I check the "always use selected program" box, then it keeps opening new blank tabs ad infinitum). Notepad opens it as a few words, URLs, and gibberish; when I try to load those URLs on IE, they cannot be decoded.

Worst of all, when I tried this on a completely different, school computer, I had the same result, so I suppose I'm doing it wrong.

I am not certain whether this problem is at all related to my use of the gadget, but the two seemed to occur simultaneously.

This is absolutely bizarre and incredibly frustrating; does anybody have any clue what is going on? Eebster the Great (talk) 03:04, 17 December 2008 (UTC)

Make sure the check box in the editing section of your preferences entitled "Use external diff by default (for experts only, needs special settings on your computer)" is unchecked. You might need to clear your cache for the changes to take effect. Graham87 06:49, 17 December 2008 (UTC)
Hey, thanks; that solved everything. I have no idea how that got checked, but it seemed to automatically happen when I enabled the gadget. Eebster the Great (talk) 03:32, 18 December 2008 (UTC)

Problem with the new File: namespace[edit]

I just noticed when look at a picture that since the {{Assessments}} template automatically links to the commons nom of the same page/file name. This causes a problem when a picture was nominated before the change, so now it links to commons:FPC/file:Image instead of commons:FPC/image:Image so you cannot get to the previous nom. Sorry if this is the wrong place to bring this up. – Jerryteps 09:11, 17 December 2008 (UTC)

Hmm, this only appears to happen on the commons page, not the wikipedia page... I think I should take this to commons but im not sure where to in commons. – Jerryteps 09:16, 17 December 2008 (UTC)

Maybe try the Commons village pump? Graham87 09:37, 17 December 2008 (UTC)


Is there a point to the three sets of includeonly's at Dewey Lemley? --Pascal666 (talk) 18:27, 17 December 2008 (UTC)

Apparently not - I deleted them and it looks OK. I suspect operator error. – ukexpat (talk) 18:40, 17 December 2008 (UTC)

Explain the file namespace?[edit]

What's with the change from "image" to "file" for media? please place a {{tb}} on m talk page when awnsered.--Ipatrol (talk) 20:01, 17 December 2008 (UTC)

There are other kind of files besides images. Garion96 (talk) 20:07, 17 December 2008 (UTC)

HTML span display error[edit]

I think I may have found a bug with the wiki software, but I wanted to ask people that probably know a little bit more about the subject than I do in case I'm doing something wrong instead of the software. On User:Dudemanfellabra/Sandbox2, I'm trying to get each column to display as inline-block. I can do this with divs (which is the current state), but this doesn't work in IE. IE only applies display:inline-block to tags that originally display inline (such as span). When I change the div of each column to a span, though, I get unexpected results. Instead of surrounding the entire column, the span is broken up and applied to each individual bit of text inside the column. After a little looking around, I came across what I think to be the culprit. The code entered into the editbox is first, followed by the actual HTML code rendered by the software (found by viewing the source of the page):

My code (with some extra line breaks so the page isn't stretched too far):

And the output code (again with some extra line breaks so the page isn't stretched too far):

As you can see, the span should surround all the inner divs, but is instead applied to every element in the column. I went a step further also and tested my code in an external editor, and it worked fine. The column was 60% width, the divs were all in the right place, etc. Back in the wiki editbox, I tried to figure out what was causing the problem, so I played around with the display tag, and found that anytime a value is set to display (in this case inline-block), the code breaks; removing the display parameter fixes the problem. I think there may be a bug with wiki software and how it interprets the span tag with a display parameter. --Dudemanfellabra (talk) 21:09, 17 December 2008 (UTC)

According to the HTML specifications, a block-level element (e.g. a DIV) may not be contained inside an inline element (e.g. a SPAN). MediaWiki can (and Wikipedia does) use a program called HTML Tidy that fixes that sort of thing in exactly the way you see above before the HTML is output to the browser. I suppose you could try replacing the divs with spans with display:block. Anomie 22:12, 17 December 2008 (UTC)
HTML Tidy wants your code to be like this:
Dendodge TalkContribs 22:27, 17 December 2008 (UTC)
I understand that an inline element is not supposed to contain block elements.. but giving span the display:inline-block command effectively turns it into a block element, which should be able to hold other block elements. Using spans with display:block doesn't fix the problem because the templates called inside the spans contain divs. Shouldn't a span that contains display:block or display:inline-block be treated just the same as a block element by HTML Tidy? --Dudemanfellabra (talk) 22:42, 17 December 2008 (UTC)
No, "display:inline-block" just says that the span should use layout constraints like a block element, but it doesn't turn the span into a block element. Lupo 12:12, 18 December 2008 (UTC)

Passing apostrophe to template[edit]

Thoughts on how to get Template:Sockpuppet category to work with usernames that include an apostrophe such as Category:Wikipedia sockpuppets of guns'nroseslover and Category:Suspected Wikipedia sockpuppets of I'll TELL You What Time It Is!? --Pascal666 (talk) 23:34, 17 December 2008 (UTC)

I think this is a (MediaWiki) bug. I've tested a number of things and it probably shouldn't be doing what it's doing. {{Nihiltres|talk|log}} 05:23, 18 December 2008 (UTC)
See bugzilla:16474. It is a common complaint that some magic words return some html entities escaped pre-comparison. For comparing a {{PAGENAME}} you need to use &#39; for an apostrophe. This is considered buggy behavior by many, as is the fact that magic words starting with list formatting characters such as an asterisk are treated as starting on a newline. {{PAGENAME}} on a page like * will start a new bullet list, for example. Did these two edits fix the template calls in question? [12] [13] -- --Splarka (rant) 08:33, 18 December 2008 (UTC)
Thank you. That does appear to have fixed the problem. I have updated the template documentation per that bug. --Pascal666 (talk) 11:38, 18 December 2008 (UTC)
  • Related: a possessive apostrophe isn't included in a link like a plural s is. Instead of George Sampson's we are presented with George Sampson's when trying to link the possessive while plurals like whales are nicely underlined and linked. - Mgm|(talk) 11:51, 18 December 2008 (UTC)
I've run into similar problems before. For example, those userboxes that say thinks like "this user is an administrator" etc, the verify link doesn't work with usernames that have apostrophes in them. There are always workarounds though. l'aquatique |Happy Hannukah!| talk 15:30, 18 December 2008 (UTC)
Which is why we have templates for {{'}} and {{'s}}, which essentially pass the HTML entities. --—— Gadget850 (Ed) talk - 15:36, 18 December 2008 (UTC)

Where do the rendered html elements come from?[edit]

Templates combination :

multi-listen start
multi-listen item
multi-listen end

all are very short and intuitive. All havn't the html LI element in their sources. But put together, an LI element was rendered into the output, inside the UL element. Just as this page: [] Could someone explain this to me please? Because I can't use these template in my own wiki site. Thank you very much. Baoshan Sheng (talk) 03:58, 16 December 2008 (UTC)

I'll venture a guess (but don't really know) that this is a result of the HTML tidy filter, see [14]. -- Rick Block (talk) 04:40, 16 December 2008 (UTC)
Thanks. But I think that's html tidy's effort just as you, but after I enable the $wgUseTidy and confirm the tidy.conf configuration, the li doesn't appear.I also notice the li element has inline styles associated with it. Html tidy may not be able to do that in my opinion.Baoshan Sheng (talk) 06:03, 16 December 2008 (UTC)
I download a tidy.exe and set the $wgTidyBin = "$IP/includes/tidy.exe". But what's the difference between the tidy come with mediawiki and the one download from [[15]] ?

Baoshan Sheng (talk) 07:18, 16 December 2008 (UTC)

Just as an aside, those templates are now deprecated, we use {{listen}} instead. I suggest you do the same. Happymelon 10:16, 19 December 2008 (UTC)

Citing different pages of a book using same ref[edit]

Here's what I'd like to do. On the first occurrence of my reference, I want to have the "base" ref which contains the ISBN, author, coauthor, etc. For example:

<ref name="lopez">{{cite book |author=Donald S. Lopez Jr 
 |title=Dictionary of American Library Biography: Second Supplement |publisher=Libraries Unlimited 
 |location=Littleton, Colo |year=2003 |isbn=1-56308-868-1 |oclc= |doi= |accessdate=}}</ref>

On proceeding uses, I'd like to use the above reference multiple times while citing different pages. So for example, we could have a (currently nonexistent) template "importref" which takes all of the already populated reference fields from "lopez", adding (or modifying) a field called "pages" to that reference and creating a new one. So for example,


This would produce a {{reflist}} item and [bracketed] inline notation for both references. Is it possible to do this using existing citation templates? This modular approach avoids having to create new references when adding/changing only one or two fields in a reference, while utilizing all of the already populated reference fields such as ISBN, author, coauthor. Spidern 16:13, 18 December 2008 (UTC)

This would be excellent. DuncanHill (talk) 16:16, 18 December 2008 (UTC)
This is not possible, although two-dimensional referencing (of which this is a manifestation) has been requested for years. Happymelon 16:36, 18 December 2008 (UTC)
It just seems rather a waste to repeat the entire citation with fields to denote a small change such as page numbers. There has to be a simple parser function that would grab the appropriate fields from a previous ref. Spidern 17:22, 18 December 2008 (UTC)
No, because references are generated after text parsing (so after ParserFunctions are resolved) but before page rendering (so before JavaScript coudld be invoked). It would require a change to the Cite.php code which, as I said, has been requested for years. If you know any PHP you'd get a lot of cookies if you could write a modification to allow this. Happymelon 18:08, 18 December 2008 (UTC)
Cite.php- can you give a link to the source code so we/I can examine the file.-ClemRutter (talk) 22:03, 18 December 2008 (UTC)
The whole of MediaWiki is available in subversion - the Cite extension is here. Happymelon 22:07, 18 December 2008 (UTC)
You could try a different approach, by having separate bibliography and reference sections. You could then cite the book fully in the bibliography, and your references would appear as <ref>Lopez, page 42</ref>, <ref>Lopez, page 83</ref>, etc. This is the way it's done in many featured articles. —  Tivedshambo  (t/c) 18:17, 18 December 2008 (UTC)
The trouble with that is that when one clicks on the numbered ref in the text, on is not then actually given the reference, one has to scroll about to find out what the reference is. DuncanHill (talk) 18:30, 18 December 2008 (UTC)
The solution intended by proponents of the 2D cite system is that the code would look like this:
 Lorem ipsum dolor sit amet,<ref name=foo name2=p22>Page 22</ref> consectetur adipisicing elit,
 sed do eiusmod tempor incididunt ut labore et dolore magna aliqua.<ref name=foo name2=p14>Page
 14</ref> Ut enim ad minim veniam, quis nostrud exercitation ullamco laboris nisi ut aliquip ex 
 ea commodo consequat.<ref name=bar name2=vol1>Volume 1</ref> Duis aute irure dolor in 
 reprehenderit in voluptate velit esse cillum dolore eu fugiat nulla pariatur.<ref name=baz>Other
 ref</ref> Excepteur sint occaecat cupidatat non proident,<ref name=bar name2=vol3>Volume 3</ref>
 sunt in culpa qui officia deserunt mollit anim id est laborum.<ref name=foo name2=p150>Page 
 <ref name=foo>Book, by Joe Bloggs</ref>
 <ref name=bar>Other book, by Other Author</ref>

And the references would output something like:

  • Book, by Joe Bloggs
    • Page 22
    • Page 14
    • Page 150
  • Other book, by Other Author
    • Volume 1
    • Volume 3
  • Other ref
As I'm sure you can see, such would be a huge improvement. Happymelon 18:28, 18 December 2008 (UTC)
Do you have any links to prior discussion about this feature? It would be useful for any potential developer to see just how much demand there is for this feature. Spidern 18:47, 18 December 2008 (UTC)

There is already a patch written that would bring this functionality: If someone properly tests it, and pushes the developers, I'm sure it can be implemented relatively quickly, now that this patch exists. --TheDJ (talkcontribs) 18:34, 18 December 2008 (UTC)

That patch is fifteen thousand code revisions out of date, DJ - I'd be astonished if it still worked. Happymelon 18:51, 18 December 2008 (UTC)
Having page number support built in to the <ref> tag seems to be the most logical way to go. Happy-melon's proposed solution seems a bit roundabout, in that there should not have to be a "name2" attribute as well as a page number inside the tag itself to circumvent the issue. Spidern 18:42, 18 December 2008 (UTC)
Yes, having a 'page' attribute solves this particular problem, but what happens if the subdivision you want is not a page? Suddenly you've backed yourself into a corner because you're committed to using this nes "page" attribute even though it's not strictly relevant. What if you're referencing different sections of a documentatary and want to cite the times? Or Clauses of a legal document, or volumes in a journal, or any one of a host of other legitimate uses of multi-dimensional referencing. Just implementing a quick hack to solve one immediate problem is very rarely the best way to proceed in the long run. Happymelon 18:51, 18 December 2008 (UTC)
In that case, how about an attribute called "append" which works like this:
<ref name="foo" append="page::28" />
<ref name="foo" append="volume::7" />
<ref name="foo" append="time::2:48" />
This makes the usage a bit more general, giving the editor the ability to append any field to the main reference. The "::" is just an example of a delimiter, anything else could be used. Spidern 18:59, 18 December 2008 (UTC)
Indeed, but it's still inflexible in that content can only be appended, when this might not always be desirable. I know I'm playing devil's advocate here but you must see my point: solutions should be designed to address the underlying problem, which is a lack of depth in the cite structure, rather than any particular symptom of that problem. Also, your proposal is essentially a macro to duplicate the majority of a reference and just 'fill in' occasional unique components. Why is that preferable to a dimensioned ref? Why is duplicating the bulk of the reference every time it is used even a good idea when the alternative is to have one main reference containing the bulk of the citation, and 'sub references' holding only the parts which are unique to each instance, all bundled together and stirred up with love by the candy man/MediaWiki parser? :D Happymelon 19:14, 18 December 2008 (UTC)
Okay, so it's still a bit of a hack but it would address a large portion of the problem (at least the one I'm targeting) which is unnecessary creation of references with repeat info. One part it does not address is the creation of multiple references at the {{reflist}}, which would be easier to target if your 2D scheme were in place. I suppose as long as I could stack fields like the following, I'd be happy:
<ref name="foo" alias="foo2">page=28</ref>
The parser could automatically figure out the type of the parent reference (foo) as being a "{{cite book}}" (depending on the case) and then append the field to the parent ref. The link would point to a "Page 28" indent of the parent reference. The "alias" attribute (or name2, if it suits your fancy) could itself be referenced in the place of "name" when doing bare references from the same page, i.e.:
<ref name="foo2" />
Spidern 19:33, 18 December 2008 (UTC)

Another approach to this—one I like and have used, but I know some people aren't as happy with it—is to put repeatedly-referenced bibliographical information in a separate "References" section using {{citation}}, then use <ref> elements that contain {{Harvnb}} references; the footnotes are gathered by {{reflist}} in a "Notes" section above the "References". Then each footnote contains page information and links successfully to the bibliographical information. See Blackface for an example of this technique. - Jmabel | Talk 19:42, 18 December 2008 (UTC)

How I work around this: I either create independent references, or, more commonly, list all the pages in a single reference and put HTML comments with the exact page number right after each use, e.g. <ref name="book">Author, Book, Publisher, Year, p. 5, 120, 511</ref><!-- page 5-->, <ref name="book" /><!--p. 511-->, ect. davidwr/(talk)/(contribs)/(e-mail) 21:53, 18 December 2008 (UTC)

I think it would be preferable to link to the main reference, rather than the individual usage of it. Anyway, it should be possible to get the right behavior without adding another ID field; but we would need to number the references, as we do now. To modify the example above:

<ref name=foo>Book, by Joe Bloggs</ref> <ref name=bar>Other book, by Other Author</ref>

And add a new syntax, i.e. <ref name=foo detail="Page 22"/> which could render something like


And the references would output something like:

  1. Book, by Joe Bloggs
    1. Page 22
    2. Page 14
    3. Page 150
  2. Other book, by Other Author
    1. Volume 1
    2. Volume 3
  3. Other ref

Steve Sanbeg (talk) 23:32, 18 December 2008 (UTC)

Would get messy if you wanted to incorporate links, templates and (especially) formatting into the "detail" field; couldn't use double quotes, for instance without escaping, which I guarrantee will never be implemented. Why is it so much harder to write <ref name=foo SomeOtherId=bar>stuff-that-should-really-be-tag-content</ref> than it is <ref name=foo SomeOtherId="stuff-that-should-really-be-tag-content"/>? Happymelon 23:43, 18 December 2008 (UTC)
The problem there is that there has to be an extra <ref name=foo> somewhere to hold the "main" reference that all the sub-refs are subordinate to (not that that's an insurmountable problem). BTW, you could use double quotes without escaping, either as {{#tag:ref||name=foo|whatever="see, quotes!"}} or <ref name=foo whatever='"see, quotes!"'>. Anomie 00:13, 19 December 2008 (UTC)
I just don't think it's necessary to list a page number twice, and would not be obvious which one is used for what. I don't think it matters to much where the text is, either my previous suggestion or something like <ref issubref name=foo>Page 22</ref>, etc. should be sufficient. It shouldn't make much difference whether templates or links, etc are included, since those should be handled after the extension has done its thing, although I realize that may be less intuitive in a parameter.-Steve Sanbeg (talk) 21:28, 19 December 2008 (UTC)

Link to Ref Desk[edit]

How do I link to a section on a Ref Desk page that is current but in such a way that it will still be valid when the day gets archived? -- SGBailey (talk) 23:33, 18 December 2008 (UTC)

If you want a wikilink to an updated version of the section then you could use #ifexist documented at Help:ParserFunctions##ifexist:. For example, write {{#ifexist: Wikipedia:Reference desk/Archives/Humanities/2008 December 17 | [[Wikipedia:Reference desk/Archives/Humanities/2008 December 17#Phenomenological life]] | [[Wikipedia:Reference desk/Humanities#Phenomenological life]] }} which renders as Wikipedia:Reference desk/Archives/Humanities/2008 December 17#Phenomenological life. This should continue to work if the page is correctly archived.
It's easier if you can accept a url to the page version at the time you make the url. Go to Wikipedia:Reference desk/Humanities and click "Permanent link" in the toolbox. Then click the section name "Phenomenological life" in the table of contents and copy the url from the browser address bar. I currently get You will get another oldid and thereby another version of the page if it has been edited since. PrimeHunter (talk) 00:54, 19 December 2008 (UTC)
That all sounds reasonable - just some more minor problems - What and where is this "toolbox" you speak of, and how can I get the link if I have content tables disabled in My preferences? Ah, I can generate it manually now I've seen the format, and I can get the oldid from the page history. -- SGBailey (talk) 07:04, 19 December 2008 (UTC)
The default MonoBook skin and some others have a toolbox in the left column of the page under a heading saying "toolbox". There may be skins with no toolbox heading and no "Permanent link" option. It is the same url you get by clicking the date of the most recent version in the page history. If a section heading contains special characters then it can be tricky to manually generate a working url to that section. But knowing that spaces should be replaced with underscores is enough in most cases. PrimeHunter (talk) 13:08, 19 December 2008 (UTC)