Jump to content

Wikipedia:Village pump (technical)/Archive 24

From Wikipedia, the free encyclopedia

nbsp inside nowiki

I've noticed that <nowiki>&nbsp;<nowiki> does not render as &nbsp;, but rather as a plain space. Has this always been this way? —Random832 20:55, 3 March 2008 (UTC)

If you check the HTML source code for the page the mediawiki software outputs, "<nowiki>&nbsp;</nowiki>" does in fact render as "&nbsp;", but not as "&amp;nbsp;". In other words, nowiki only prevents the mediawiki software from doing strange things to to the non-breaking space before it sends it to your browser. Your browser is the software doing the strange thing to "&nbsp;" and turning it into a (non-breaking) space. Of course, this is not too strange, but rather just part of html. JackSchmidt (talk) 21:06, 3 March 2008 (UTC)
Yes, it's always been that way. You have to use &amp; to get &action=view to work like that. --MZMcBride (talk) 06:26, 4 March 2008 (UTC)
Why doesn't <nowiki> do the same to html entities as it does to html tags? that is, why doesn't it change & to &amp; like it changes < to &lt; in <span>? —Random832 15:14, 4 March 2008 (UTC)

Problem with posting at Wikipedia talk:Mirrors and forks

I just tried to post some plain text at that page (with nothing even resembling a URL), as a test, and got an error message upon trying to save:

Spam filter notice

The following link has triggered our spam protection filter: h-t-t-p-:-/-/-w-w-w-.-g-o-v---c-e-r-t-i-f-i-c-a-t-e-s-.-c-o-.-u-k [Note: hyphens added to avoid triggering the filter when posting here.]

Either that exact link, or a portion of it (typically the root domain name) is currently blacklisted.

I have no idea where that is coming from; could someone else take a look? -- John Broughton (♫♫) 15:09, 4 March 2008 (UTC)

The problem is that the spam blacklist triggers if a link appears ANYWHERE in the page (even if it's not the section you're editing) - and the developers don't seem to know why this is stupid. —Random832 15:16, 4 March 2008 (UTC)
Ah, thank you. And thank you for breaking the link so that editors can now actually post to that page. -- John Broughton (♫♫) 16:57, 4 March 2008 (UTC)
The message is from MediaWiki:Spamprotectionmatch and you are not the first to be confused. I suggest explaining that the link is not necessarily in the edited part of the page. PrimeHunter (talk) 17:38, 4 March 2008 (UTC)

Call for MediaWiki hackers

From Brion:

The last couple of years we’ve had limited success with [Google Summer of Code], in part because we’ve been so shorthanded on mentors that we can’t support more than one or two students. I’m looking for a few MediaWiki hackers who’d like to help out this time around.

If you're interested, please see Brion's web posting.

For background:

-- John Broughton (♫♫) 17:25, 4 March 2008 (UTC)

wantedcategories

Special:Wantedcategories only goes to 1000, which means it omits all categories that have only one call to them (the ones most likely to be easily fixable typos). Is there any chance this query could be altered to allow more entries to be shown? —Random832 20:01, 4 March 2008 (UTC)

Most Special: pages are limited to the first 1,000 entries. Some have been cached for years, others update once every three days. The easiest way to get a full list is to use the Toolserver. There may be a tool already written, but probably not. You can use the tswiki:Query service to request DB queries. Cheers. --MZMcBride (talk) 21:03, 4 March 2008 (UTC)
Correction: there are "live" and "cached" Special pages, and most cached ones are limited to 5000 entries, see m:Help:Special page. —AlexSm 22:35, 4 March 2008 (UTC)

I ran the query for you. Results are on JIRA (a little big to put on a wiki page). Let me know if you have any trouble/need another run. - AWeenieMan (talk) 01:31, 5 March 2008 (UTC)

Wow 26,684 entries. Would there be anyway to bot-notify the talk-page of the pages they appear on to check for a typo or create a categoy? This makes the 4,000 user pages without registered users look small. MBisanz talk 02:43, 5 March 2008 (UTC)
Definitely possible, but I think we are a ways away from that at the moment. For example, there are currently 2366 redlinked categories of the format "Suspected Wikipedia sockpuppets of Example." Now it would seem silly to tag such talk pages, because a chunk like this could benefit from one decision being made for all of them (should they be created and put into Category:Suspected Wikipedia sockpuppets, should they all be left redlinked, or should they be depopulated (and admins trained on when to create them) the {{sockpuppet}} template be altered not to create them by default?). This is just an example of some of the category work that needs to be done beforehand. - AWeenieMan (talk) 05:06, 5 March 2008 (UTC)

Empty Watchlist?

Is there another way to empty your watchlist? Evidently, when programming my bot, I didn't do something right because my bot now has 52,700 articles on its watchlist. I've tried going to Special:Watchlist/raw and simply deleting them all, but the request times out every time. Is there another way? -- SatyrBot (talk) 15:33, 4 March 2008 (UTC)

Heh, I just found this, oh dear... - perhaps a developer needs to fix it manually.  :-/ x42bn6 Talk Mess 17:06, 4 March 2008 (UTC)
There is also Special:Watchlist/edit, but I suppose that might time out too... Arthena(talk) 17:08, 4 March 2008 (UTC)
That page doesn't even load because there are too many :) How does one go about asking a developer to clear it out manually? -- SatyrBot (talk) 18:29, 4 March 2008 (UTC)

Try this: open Special:Watchlist/raw from your other account (and save the list externally just in case). Open "log out" link in a new browser window, then sign in as a bot account. Switch back to the raw watchlist window, press "Update Watchlist" button, it should update the list for your bot account. I just tried it and it worked for me. P.S. Please do not participate in discussions from your bot account. —AlexSm 18:51, 4 March 2008 (UTC)

It surprises me that that would work as this form appears to, for potential security reasons, contain a token. Perhaps this is ignored when the form data is received by the server? (probably not good either). — CharlotteWebb 20:17, 4 March 2008 (UTC)
I don't see any security concerns at the moment, I guess token is a token no matter which account generated it. If a token was linked to user account, then I think a token from edit page (or API generated) would work, and then a little userscript to generate and submit Watchlist/raw form would do the job. —AlexSm 23:01, 4 March 2008 (UTC)
Sorry, I didn't see this method. This seems far superior as a one-off fix for a crazy problem (than the api.php/index.php nonsense I suggest below). The clear button would be nice for casual users of course. JackSchmidt (talk) 01:20, 5 March 2008 (UTC)

I believe there used to be a Special:Watchlist/clear page containing just one button to clear the watchlist (without listing all of the pages). Apparently this now redirects to Special:Watchlist/raw, but I'm not sure this is ideal. — CharlotteWebb 19:20, 4 March 2008 (UTC)

Perhaps your best bet is to get your bot's watchlist from the API, and unwatch a few (tens of) thousand(s of) pages by loading the html http://en.wikipedia.org/index.php?title=THEPAGE&action=unwatch from a script. And perhaps a dev ought to add a hook so that you can't make your watchlist so long that you can't load it (or just put a maximum limit on the number returned from Special:Watchlist/raw). Happymelon 20:13, 4 March 2008 (UTC)
You can also use http://en.wikipedia.org/w/index.php?action=ajax&rs=wfAjaxWatch&rsargs[]=THEPAGE&rsargs[]=u to unwatch pages without getting a full page back, which should lessen the server load a little. If you need to unwatch a few thousand articles, check out User:JackSchmidt/JS_Watchlist.js which is designed for adding a few thousand articles to the watchlist. You should be able to change
["rsargs[]","w"]
to
["rsargs[]","u"]
to do the unwatch. The code does 1 AJAX hit per article, but maintains a queue so that only 1 request is active to the webserver at a time. JackSchmidt (talk) 20:58, 4 March 2008 (UTC)
Still, a lot of queries to server, and I don't see API query to get one's watchlist anyway. —AlexSm 23:01, 4 March 2008 (UTC)
http://en.wikipedia.org/w/api.php?action=query&list=watchlist&wlprop=title should work well. The bot would only need 6 api queries and 50,000 almost-no-output index.php queries (but each one doing a database update). Certainly a simple clear page would be better, but I don't think a one-time use of this method would be too troublesome. JackSchmidt (talk) 01:15, 5 March 2008 (UTC)
1) action=query&list=watchlist gives you recently changed watched pages, so it has to be filtered for duplicates (when both base and talk pages are listed) and I don't think it can give you pages that were not edited in last 30 days. It can certainly be used to unwatch a lot of pages, just not all of them. 2) To mass-unwatch pages I would create a POST request imitating Watchlist/edit form. —AlexSm 03:58, 5 March 2008 (UTC)

Special:Watchlist/clear can easily be added back, see bug 13250. Oysterguitarist 00:35, 5 March 2008 (UTC)

I have a script for doing just that, at one point BCBots watchlist had 80k items. if you can get on IRC,(freenode) I can walk you through clearing your watchlist. :P (PS give me 6 hours before getting on)βcommand 15:42, 5 March 2008 (UTC)
Thanks for the responses. I currently have a script running, but at 5 seconds per item, it's going to take nearly three days of continuous running. Other options? Beta, what channel? -- SatyrTN (talk / contribs) 21:28, 5 March 2008 (UTC)
Hmm, 5th post in this section, which describes how to clear your watchlist in less than a minute? —AlexSm 21:34, 5 March 2008 (UTC)
Hi, Alex - I tried that one, too, and it doesn't work. The call to clear the database is the same, no matter whether I started logged in as me or as the bot. The only option I haven't tried yet is the POST imitating the edit form, which I may have to try next :) -- SatyrTN (talk / contribs) 22:04, 5 March 2008 (UTC)
SatyrTN irc://freenode/wikimedia-tech is a good place to find me. βcommand 22:55, 5 March 2008 (UTC)

Articles created by a user: is there a list?

This might be a ridiculously simple question, but: Is there a way to list the pages created by a given user? 52 Pickup (deal) 15:51, 4 March 2008 (UTC)

Actually a fairly frequent question. Yes, there is a tool - here. May run a bit slowly (or not; someone was looking at that in the last couple of days). -- John Broughton (♫♫) 17:00, 4 March 2008 (UTC)
Just what I was looking for. Thanks! 52 Pickup (deal) 18:30, 4 March 2008 (UTC)
Yet it seems not to be working. According to the tool in question, either I do not exist or I have no edits. Although it is true that I have never created a mainspace page (with this account), the returning message is not something that really makes sense. I tried with a couple of other editors, including Mr Broughton here, and the result was the same in every case. With the "English" selection having already been made, I cannot find what on Earth I could possibly have done wrong. Unless, of course, someone's been doctoring the edit records, in which case our problems are way worse than a simple tool glitch. And the plot thickens... Waltham, The Duke of 15:25, 5 March 2008 (UTC)
try [this βcommand 15:36, 5 March 2008 (UTC)
When I ran it, it only listed mainspace creations (including redirects). Since that was all I was interested in at the time, I wasn't so worried. If the tool can be modified to cover other spaces, that would be nice too, I guess. 52 Pickup (deal) 15:33, 5 March 2008 (UTC)
Betacommand's link is an interesting one; although it's not exactly relevant (page creations are not mentioned), it provides links to every single edit. That's novel—for me, at least.
As far as the initial link is concerned, the tool's problem must be mostly related to my rather slow Internet connection; 52 Pickup appears to have no mainspace creations either. (Hey, now that you mention redirects, I realise that I actually have plenty of mainspace creations, therefore I ought to see pages appear for me as well. In a perfect world, that is...) Waltham, The Duke of 17:12, 5 March 2008 (UTC)
Look near the bottom of the page, it gives a list of non-talk, non-user, non-redirect pages you have created. βcommand 17:16, 5 March 2008 (UTC)
I hadn't bothered to go that far down... This is most useful. Thank you very much, βcommand. Waltham, The Duke of 22:41, 5 March 2008 (UTC)

Someone semiprotecting a page made it dissapear from my watchlist

The userpage User talk:Alison was protected and then stop appearing in the recent changes of my watchlist. ViridaeTalk 23:13, 4 March 2008 (UTC)

As with moves above, I think protections don't usually show up until someone edits on them again. At least that is what I have had since I have been on Wikipedia. Is there a bug open for this? Woody (talk) 23:16, 4 March 2008 (UTC)
Bugs: 5546 881 3067 3465 4898 5912 7181 7790 8804 8969 9189 9452 9689 10986 11750 12046 12062 12375 12376. --Splarka (rant) 09:25, 5 March 2008 (UTC)
Do you mean that all edits (also editing the text) disappeared from your watchlist? PrimeHunter (talk) 01:01, 5 March 2008 (UTC)
If you don't use the expanded watchlist, you only see the last edit. If that last edit is a log event (move, protection), the page is not listed in the watchlist until a subsequent edit is made. Gimmetrow 10:04, 5 March 2008 (UTC)

New pages

What has suddenly gone wrong with Special:Newpages? I click 500 or 250 and it shows me only 50. I edited the URL directly to set a limit of 500 and it showed only 50. (I just tried 20 and 100 and they set the right limit. 250 and 500 set the URL to request an appropriate limit but the page still displays 50.)

My normal practice on New pages patrol is to hide patrolled edits, view 500, then start at the bottom, which usually give me pages created early in the day so they've had plenty of time to revise it. Right now I can't do that but I suppose I could edit the URL to an offset of 500 and the limit of 50 would be okay. ... Yep, that works. Sbowers3 (talk) 05:29, 5 March 2008 (UTC)

It looks like the highest number it can display is 200. Setting the URL to 201 makes it display only 50. – FISDOF9 05:42, 5 March 2008 (UTC)
That was done because somebody used an exploit relating to the page setting of Special:Newpages to take down the Wikipedia servers. -- Prince Kassad (talk) 06:23, 5 March 2008 (UTC)
Specifically, this is one of the special pages that can be transcluded, and someone added {{Special:Newpages/5000}} to German WP's main page (through an unprotected template). --Dapeteばか 11:37, 5 March 2008 (UTC)

Glitch

Does anybody see the stray text at the top of Rathcoole, Dublin? Where is it coming from? (I'm using Firefox, if that makes any difference.) Paddy Simcox (talk) 07:54, 5 March 2008 (UTC)

It was from a missing right brace ("}") in this edit to Template:Expand. I've fixed the problem - if it still happens on an article you're viewing, purge it. Graham87 08:21, 5 March 2008 (UTC)
Thanks, it's gone now. Paddy Simcox (talk) 08:42, 5 March 2008 (UTC)

Edits do not show up in IE (NOT CACHE problem)

In the article Free electron laser#Medical applications, in the Medical applications section, I added two paragraphs above the paragraph starting with "At the 2006 annual meeting of the American Society for Laser..." This shows up fine in Firefox on Linux, Safari and Firefox on Mac, but not IE6 on Win2000 Pro (and other IE's it appears). However, from IE6 if I click on Edits for that section I see my two paragraphs fine. Quintuple checked about the cache clearing, so I do not believe that is the problem. Gabella (talk) 16:30, 5 March 2008 (UTC)

Watchlist view in IE8

The watchlist is kinda rendering funny, spiking up the cpu and stuff. (See here: http://cid-b24d2193d45b2df2.skydrive.live.com/self.aspx/Public/iWiki.PNG ) Could someone take a look at what is breaking the page, and fixing / hacking whats necessary. Or at least use the IE7 rendering mode (for that specific page , using <meta http-equiv="X-UA-Compatible" content="IE=7" />) so that it becomes at least usable. --soum talk 05:16, 6 March 2008 (UTC)

Adding more features

The currently available features are not sufficient to meet my needs. E.g., I want to implement proxy designation acceptance that can be done with two clicks, and the inputbox feature does not give me the flexibility to do that, because it won't take templates and such as parameters. Also, as mentioned above, I need to use {{CURRENTUSER}}, but it's not enabled. Bugzilla says, "Note that voting is nowhere near as effective as providing a fix yourself." How do I go about doing that? I know some Java, but is that applicable? Am I allowed to do it myself or do I need permission? Thanks, Ron Duvall (talk) 02:34, 22 February 2008 (UTC)

The feature request in question is marked WONTFIX. It doesn't matter if you implement it yourself, it still won't be put on Wikipedia. Think of another way to do what you want to do. For instance, the username is available to JavaScript as wgUserName. -- Tim Starling (talk) 03:34, 22 February 2008 (UTC)
If I read the use-case correctly, I think the idea is that Joe Schmoe user would be doing the two clicks, probably only a few times ever. Am I right in thinking that javascript is mostly helpful for a single user to add new features only for their own account, so is not likely to be useful to all users by default?
I suppose something could be added to some of the site's default javascript to change all a.href values to replace "User:CURRENTUSER" with "User:" + wgUserName, but where does one get consensus to do that sort of thing? JackSchmidt (talk) 03:55, 22 February 2008 (UTC)

But they have it on Uncyclopedia! See http://uncyclopedia.org/wiki/Template:USERNAME . Are we going to let those guys one-up us? Ron Duvall (talk) 05:00, 22 February 2008 (UTC)

I wrote the original (and a few followups) implementation of {{USERNAME}} on Uncyclopedia (which is now used on much of Wikia). It makes sense for Uncyclopedia, as they had 35% of all page views from logged in users (Wikipedia has like 1% or less?), and they use it for practical joking (not needed here). It wouldn't really work for much anyways, you couldn't use it for content modification, ie via ParserFunctions. If you need to link to a user's page, use Special:Mypage, eg Special:Mypage/monobook.js. If you need to talk to a user in wikitext, say "hey you!". If you need to control content based on the user's username, write the appropriate JS function and propose it. Note that {{subst:USERNAME}} is possibly going to be implemented someday (as this won't break caching). --Splarka (rant) 05:20, 22 February 2008 (UTC)

HIDDENCAT

A new feature has gone live on Wikimedia wikis: __HIDDENCAT__ . Adding this magic word to a Category: description page hides the category from all of the pages in that category. So if it were added to Category:Wikipedia and the Main Page were categorized in Category:Wikipedia, the category description page would list the Main Page, however, the Main Page wouldn't show the category at the bottom of the page, all without any nasty CSS hacks.

The magic word has the potential to help with maintenance categories, stub categories, and more. --MZMcBride (talk) 06:07, 22 February 2008 (UTC)

Neat feature, it should unclutter a good number of articles. Can it be implmented selectivly. Like for Category:Wikipedia, could it not show on the Main Page, but show on the page WP:AFD? MBisanz talk 06:19, 22 February 2008 (UTC)
No, not with this syntax. __HIDDENCAT__ should only be put on categories that you want to be hidden from all articles. -- Tim Starling (talk) 08:32, 22 February 2008 (UTC)
What about Category:Living people? I think it would be useful there. Waltham, The Duke of 16:30, 22 February 2008 (UTC)
This would be undesirable. Content categories should always show up for viewing. How else can editors know whether an article or category has been placed or not placed in their proper categories. HIDDENCAT should only be used for admin categories. Use of HIDDENCAT in stub cats need more discussion. Are stub cats administrative only? Hmains (talk) 04:12, 24 February 2008 (UTC)
Nice. Is there a reverse feature where something can be placed on a page so it shows categories normally at the bottom, but the page is not displayed on the category pages? It could be useful for userfied articles and private sandbox drafts, but there would probably have to be a system so at least deletion categories could list the page. And while speaking of category manipulation, I miss a feature to say at a transclusion that the transcluded page should not be allowed to add categories (e.g. when showing what a template looks like). This can currently be achieved by adding a parameter to the transcluded page and code it to omit categories when the parameter is there, but that's often non-trivial. PrimeHunter (talk) 15:57, 22 February 2008 (UTC)
I thought this feature was discussed and rejected as being too prone to vandalism? We need to be able to generate a list of hidden categories to check that the feature is not being abused. Incidentally, if you can generate a list of all pages using the __HIDDENCAT__ magic word, could I repeat my oft-asked plea for a way to list pages that are using the DEFAULTSORT magic word? Carcharoth (talk) 03:32, 23 February 2008 (UTC)
Without WP:BEANS, should there be a centralized discussion on what categories and types of categories should be hidden? MBisanz talk 04:13, 23 February 2008 (UTC)

I've just been bold and added it to Template:MonthlyCleanupCat It should eventually affect all the categories that use that template. Graham87 13:02, 23 February 2008 (UTC)

One thing I just noticed is that the parent cat will not show up in sub-cats. This is not desirable. __HIDDENCAT__ should not affect Category space. --Farix (Talk) 18:55, 23 February 2008 (UTC)
Do you have examples? Hmains (talk) 04:12, 24 February 2008 (UTC)
I don't think it affects the display of pages in the Category namespace. It absolutely does not on a default install of current mediawiki. JackSchmidt (talk) 06:15, 24 February 2008 (UTC)
For example, Category:Articles lacking sources from February 2008 does not show that it is a subcategory of Category:Articles lacking sources.--Patrick (talk) 10:48, 24 February 2008 (UTC)
This fact seems to be quite ok; they are administrative categories, not main space categories. Hmains (talk) 18:21, 24 February 2008 (UTC)
So how do you navigate from Category:Articles lacking sources from February 2008 to Category:Articles lacking sources? Carcharoth (talk) 22:16, 24 February 2008 (UTC)
I have edited {{MonthlyCleanupCat}} to include a link to the parent category. Tra (Talk) 00:23, 25 February 2008 (UTC)
You are correct, my mistake. I created a patch and posted it to bugzilla:13140 (which has now been fixed). JackSchmidt (talk) 00:14, 25 February 2008 (UTC)
Thanks! Any idea how to answer my question: How many categories has this tag been placed on? Carcharoth (talk) 03:16, 25 February 2008 (UTC)
I put HIDDENCAT on Category:Bridges to Rockaway, Queens and it does appear in the parents, and the parents appear on it. -- SamuelWantman 10:27, 25 February 2008 (UTC)
Fixed in revision 31250 by huji. Parent categories now show on category pages regardless of __HIDDENCAT__. JackSchmidt (talk) 13:15, 25 February 2008 (UTC)

It would nice if the HIDDENCAT would show visually on the category in which it is placed. As things are now, you always have to edit the [admin] category to see whether HIDDENCAT word exists where you want it. Hmains (talk) 04:12, 24 February 2008 (UTC)

I made {{hiddencat}} putting the tag and giving the message.--Patrick (talk) 10:57, 24 February 2008 (UTC)
That's a nice idea, but we will simply end up with the same situation as we had with {{DEFAULTSORT}}. Why use a template instead of a magic word? Carcharoth (talk) 22:16, 24 February 2008 (UTC)
As long as the use of HIDDENCAT is not visible on the rendered category page, this is a workaround.--Patrick (talk) 00:51, 25 February 2008 (UTC)
I see that this edit has replaced the template with the magic word, thus removing the wording you has put in that template. Carcharoth (talk) 03:15, 25 February 2008 (UTC)

How many categories has this tag been placed on? - until we have an automagically generated list of the categories that this magic word has been used on, it shouldn't be used. Period. Otherwise it will be used for vandalism. Carcharoth (talk) 22:10, 24 February 2008 (UTC)

I think discussion about using HIDDENCAT should happen at WT:CAT. I've started a discussion there related to using this on all category intersections and repopulating their parents. I understand Carcharoth's concern about this being a source of potential vandalism. I'd like to see some well defined guidelines on using this feature, and perhaps a bot to monitor its use. I think it is a wonderful addition that solves many problems (maintenance categories, category clutter, multiple intersections). One way a bot could patrol this is to require that hidden cats be "registered" by adding them to a protected list. That way, a bot could monitor recent changes to category pages and see if any unauthorized changes were made. -- SamuelWantman 10:27, 25 February 2008 (UTC)

Sorry, but I don't really understand how or why this feature could or would be used for vandalism. Would someone like to carefully explain? Happymelon 11:38, 25 February 2008 (UTC)
Adding hidden categories that are inappropriate. If this gets missed in the initial addition, people won't spot the category tag at the bottom of the page because, well, it won't be there. Also, adding HIDDEN CAT to a category to "quietly destroy it", and nobody else noticing that the category tag has disappeared from a group of articles. I know this violates WP:BEANS, but magic word vandalism is nothing new. What is new is having such a powerful magic word to change the way things look. Carcharoth (talk) 13:28, 25 February 2008 (UTC)

Still concerned, but if carefully monitored, and if it is a useful step towards category intersection, then I'll stop objecting and worrying about vandalism. Carcharoth (talk) 13:36, 25 February 2008 (UTC)

There should be a way to see hidden categories on an article if you want to - either through a preference, or through the categories being rendered but hidden by default with CSS and can be overridden in user CSS. —Random832 17:47, 25 February 2008 (UTC)

This is being discussed at wikitech-l. User:Simetrical gave sort of a blanket no, already at [1]. I wrote some simple javascript to show the hidden categories on a line below the normal categories line. If people think this is useful, I can wrap it up for others to use. JackSchmidt (talk) 18:05, 25 February 2008 (UTC)
Then it needs to be made available in a javascript variable along with all the wg stuff - fetching and parsing the wikitext is an unacceptable solution, particularly since many categories are added by templates. —Random832 18:16, 25 February 2008 (UTC)
I just use mw:API, http://en.wikipedia.org/w/api.php in my script. My script needs to be fixed up to handle internationalization, add some comments, maybe factor the major pieces into their own functions. It uses a helper script, and I haven't gotten around to figuring out a smooth way to cut down on load time and number of loaded pages. I was going to ask Lupin about his scripts at some point. At any rate, if there is interest I'll clean it up, but it works for me and my friends, so I'm happy with it right now. JackSchmidt (talk) 18:44, 25 February 2008 (UTC)

Is there a way to force a category to be displayed? Perhaps in a user's personal settings? -- Ned Scott 06:45, 26 February 2008 (UTC)

Just tick the box at My Preferences -> Misc -> Show hidden categories. —AlanBarrett 07:15, 6 March 2008 (UTC)

See rev:31262 and rev:31257 for new features related to HIDDENCAT. The version of MediaWiki that Wikimedia wikis are currently running is 1.44.0-wmf.5 (d64f667). --MZMcBride (talk) 07:20, 26 February 2008 (UTC)

Thanks! Those are great changes. This is all so far-reaching it needs to go in something llike the Signpost and other places as well. I'm copying out the details from those changes you linked:
  • Put both hidden categories and normal categories into the view page HTML, but with hidden categories being unconditionally hidden with CSS. A JS show/hide toggle can be added in user/site JS.
  • Add user preference to always show hidden categories
  • Add all hidden categories to Category:Hidden categories, localised by hidden-category-category
  • Add wgVariantArticlePath and wgActionPaths to the JS variables script, needed to determine title from link href.
  • For pages with hidden categories, add a list of these categories down the bottom of the "edit" page - underneath the list of included templates.
The Category:Hidden categories was essential, in my opinion, and will need to be carefully watched. Any way to see "recent changes" on what was added to that category? Carcharoth (talk) 12:16, 26 February 2008 (UTC)
Special:Recentchangeslinked/Category:Hidden categories sort of does what you want, as it shows all the changes to all the hidden categories. It does not however show any changes to templates or pages in other namespaces that use __HIDDENCAT__. Graham87 14:27, 26 February 2008 (UTC)
I'm really looking to detect edits that add and remove "HIDDENCAT". I don't think your suggestion does that, but thanks as I'd forgotten that option. Carcharoth (talk) 15:03, 26 February 2008 (UTC)

Can someone explain the objection the devs have to hiding it with CSS? It is in no way equivalent to hiddenstructure. It's being hidden as a stylistic choice, not because it's empty structure being shoved out of the way; and the page would make perfect sense with them not hidden. siteSub (that's "From Wikipedia, the free encyclopedia", if you're wondering) is hidden by default in mediawiki. —Random832 14:29, 26 February 2008 (UTC)

Incidentally, there is a discussion underway at Wikipedia talk:Categorization#Hidden categories concerning what kinds of categories should be hidden. For the moment it is proposed that hiding be applied to all categories which classify the article rather than the article subject (i.e. maintenance cats, stub cats, "Spoken articles" etc.) Please weigh in. --Kotniski (talk) 08:37, 27 February 2008 (UTC)

There is a proposal to restructure how we display and populate categories that uses HIDDENCAT as well at WT:CAT. --SamuelWantman 09:29, 27 February 2008 (UTC)

I have now made hiddencat optional for {{MonthlyCleanupCat}} - I will make it NAMESPACE specific (unless someone else does it first) and I will fix-up all the cats that use it to implement it through the argument call and not doubly. With wishing to WP:OWN... someone could have asked me for input? Rich Farmbrough, 10:11 5 March 2008 (GMT).

TFD

FWIW, {{hiddencat}} is currently at TFD: Wikipedia:Templates for deletion/Log/2008 February 28#Template:Hiddencat. szyslak 02:21, 29 February 2008 (UTC)

If hiddencat isn't working for you, it's probably because the code in the {{hiddencat}} template was commented out for a while. It's been fixed and should work fine now. If you're still having problems, try bypassing your browser's cache. szyslak 07:25, 29 February 2008 (UTC)

Image linking to other pages

Hello, I wasn't sure where to ask this but techical seemed to apply the most. I was wondering how does one go about posting an image that links you to a different wiki page when you click on it instead of sending you to the image page? I tried a few things like [[randomwikipage|[[image:randomimage]]]], but nothing seemed to really work. Thanks for the help. Cbale2000 (talk) 17:45, 5 March 2008 (UTC)

See mw:Extension:ImageMap. —Random832 19:06, 5 March 2008 (UTC)
Is there any other way to go about doing it? I ask because I'm trying to add the function to a different wiki I don't think has the ImageMap extension. Anything simple that just makes the entire image link to some other location? Cbale2000 (talk) 05:51, 6 March 2008 (UTC)
mw:Help:Linked_images --Splarka (rant) 08:29, 6 March 2008 (UTC)

Background color in {{test6}} block message spills into next section

Why does the background color in this post-block message using {{test6}} spill into the next section, and how can it be prevented? - Neparis (talk) 04:11, 6 March 2008 (UTC)

It would appear that when NawlinWiki placed the block message a while back, the message did not have a closing </div> tag [2]. I added the closing tag and it looks fine now. - AWeenieMan (talk) 04:34, 6 March 2008 (UTC)
Thanks for fixing it. - Neparis (talk) 20:58, 6 March 2008 (UTC)

Script wanted for measuring article sizes within a cat

Hello I was looking for a script that can go through a list of articles, or a category, and calculate the mean and median article size of the articles on the list or cat. Specifically, I was wanting a program so that it could look at a cat of Start/B/Stub class articles, eg Category:Start-Class_Vietnam_articles, and then calculate the mean and median size of the corresponding articles (not the talk page that is tagged) in this category. Note that Dr pda (talk · contribs) had created a similar thing at User_talk:Dr_pda/generatestats.js which may be used as a model; his script works only on articles that have transcluded a given template, rather than having membership of a category. Thanks. Blnguyen (vote in the photo straw poll) 07:11, 6 March 2008 (UTC)

Unfortunately Dr pda is no longer active. Blnguyen (vote in the photo straw poll) 07:12, 6 March 2008 (UTC)
Here is a dinky test-version: User:Splarka/genaverage.js. You can try it with importScript('User:Splarka/genaverage.js'); (purge cache, and then go to MediaWiki:Generatestats).
Example

An example, this:

Foo
Bar
Baz
Nosuchpagelalala
Traffic lights
generates :
Number of counted pages: 3
   mean = 2870.33 bytes
 median = 3969 bytes

Counted pages: 
 * [[Bar]] (4133 bytes)
 * [[Baz]] (509 bytes)
 * [[Foo]] (3969 bytes)

Uncounted pages: 
 * [[Nosuchpagelalala]] (no such page)
 * [[Traffic lights]] (too old)
--Splarka (rant) 10:33, 6 March 2008 (UTC)

Image

I cannot insert the Image:Meda-Stemma-New.jpg in the Meda (MI) page, as coat of arms. The new link is correctly written, but in the template there is still the old image. Thanks, --DoppioM 14:25, 6 March 2008 (UTC)

Purge the server's cache, then bypass your browser's cache. That solves most issues. Stifle (talk) 14:33, 6 March 2008 (UTC)

Search engine

I have had a curious problem with the search engine interface in Firefox (that little box to the right of the address bar). I added Wikipedia some time ago and it was added as Wikipedia (English) to the box. Recently the drop-down menu turned blue (to signal the possibility of adding a search engine) and it suggested adding Wikipedia (en), which I did and removed Wikipedia (English).

Now when I view a few pages (Reading (disambiguation) being one of them), the box turns blue again and suggests adding Wikipedia (English) again.

Seems like a technical error with the server (I have no idea how it informs Firefox of the search function, nor of how it works), but it could just as easily be at my end.

If you need more information about this please drop me a line at my talk page as I am likely to forget about this and not come back. Stifle (talk) 14:33, 6 March 2008 (UTC)

Main Page Down!

The Main Page is messed up. The navigation bar to the right is missing as are the discussion, view source, and history links at the top. It can be accessed properly at Main page (different capitalization), but the "English" link on http://wikipedia.org/ goes to the messed up "Main Page". —137.186.165.215 (talk) 23:03, 6 March 2008 (UTC)

I don't see any problems. Try bypassing your cache. PrimeHunter (talk) 23:08, 6 March 2008 (UTC)
Never mind. Sorry. —137.186.165.215 (talk) 23:15, 6 March 2008 (UTC)

Change "permission error" to Unauthorized

When a user without sufficient rights tries to preform an action they're not allowed to, the message Permission Error, is displayed. I think it should just say Unauthorized since permission error sounds kind of awkward. This is just a minor change so I didn't think it needed consensus for it to happen, but apparently so. I realize this is a dull topic so if no one objects in a week or so, I'll rerequest it to be changed. -- penubag  (talk) 03:19, 1 March 2008 (UTC)

While "Permission error" certainly sounds awkward, it is a wording that is very clear about the nature of the problem in terms of the way that the system handles it. Nihiltres{t.l} 19:48, 2 March 2008 (UTC)
I'd express support for one side or other of this discussion, but I don't understand when that message is actually displayed. If an editor can't actually edit a page, then there is no "edit" tab; it's replaced with a "view source" tab. I was able (by attempting to create what is a salted page) to get a screen that is titled "Permission error", but that's not exactly a problem, because the text on the screen (This page is currently protected, and can be edited only by administrators. ... Some templates and site interface pages are ...' ) is pretty good at explaining the issue. -- John Broughton (♫♫) 22:32, 2 March 2008 (UTC)
The message usually comes up when someone uses a special page which they don't have permission to use. To give some examples, the message comes up when a non-admin tries to use Special:Blockip, a non-bureaucrat tries to use Special:Renameuser, or someone without oversight permissions tries to use Special:Oversight. I think "unauthorised" is a less technical term and is less awkward - signs saying "no unauthorised access" are common whereas signs saying "people with permission errors not allowed" are uncommon. Graham87 02:36, 3 March 2008 (UTC)
Yes, that is where it usually appears. -- penubag  (talk) 02:50, 3 March 2008 (UTC)
Thank's for the example. I tried Special:Blockip, and got "Permission error" as the title of the page, with text that said "The action you have requested is limited to users in the group Administrators." Seems fairly clear to me. It might be slightly clearer if the page was titled "Access restriction", but if there is no consensus on this, I don't think it's a particularly important problem. -- John Broughton (♫♫) 16:21, 3 March 2008 (UTC)
I personally think it would be a good idea to reword it into a more user friendly terms. Permissions in a computer sense is confusing to people without understanding. Now, as stated above, signs stating no unauthorized access or something of the like is fairly common place in my opinion and much easier to understand.Chrislk02 Chris Kreider 15:25, 4 March 2008 (UTC)
"Permission error" makes a lot more common sense than "unauthorized access". What is an "access" anyway when you are sitting still at a computer, not walking around some unfamiliar part of town? - Neparis (talk) 03:16, 6 March 2008 (UTC)
No, the title would just only display "Unauthorized", not "unauthorized access". I'd be like permission error. -- penubag  (talk)

So, can one of you admins here please make the changes, there seems to be enough support? -- penubag  (talk) 03:33, 6 March 2008 (UTC)

There is no consensus for such a change. - Neparis (talk) 21:03, 6 March 2008 (UTC)
I did it anyway, the best way to see if people don't like a change is whether they complain en masse when it is done. Since it is so minor, WP:BOLD is the way to go. Prodego talk 21:08, 6 March 2008 (UTC)
Thank you Prodego. It gets irritating how every darn little change needs some mass consensus. -- penubag  (talk) 01:53, 7 March 2008 (UTC)
I don't like it, but my objection is already recorded. Happymelon 12:21, 7 March 2008 (UTC)
Yeah, it should have been a big bold red ACCESS DENIED!!! :) EdokterTalk 12:29, 7 March 2008 (UTC)

Could a preference be added to disable the viewing of anything listed under badimages?

This seems useful for obvious reasons. That way if someone doesn't want to view any number of disgusting things, they don't have to. Jtrainor (talk) 07:40, 1 March 2008 (UTC)

You could file a request at Bugzilla. It could probably be done with Javascript as well, but it would probably slow down browsing Wikipedia as it would have to load the badimages page every time you visit a page and then compare all the images on that with the images on the page. Mr.Z-man 06:10, 2 March 2008 (UTC)
Couldn't it make a daily cache of the badimages page? αѕєηιηє t/c 09:45, 2 March 2008 (UTC)
Okay, I might have come up with a more efficient way to do it with the API. Mr.Z-man 20:02, 2 March 2008 (UTC)
Script done. See User:Mr.Z-man/badimages. I've only tested it in the monobook skin with Firefox/2.0.0.12, and I make no guarantees it will work in all browsers (especially Internet Explorer) or other skins. Since it needs the whole page to be loaded before it can run, you should put it on the top of your user JS page so it loads first. Mr.Z-man 22:38, 2 March 2008 (UTC)
Pardon me for parroting my good friend Jtrainor, but this is a remarkably intelligent idea. It gives Wikipedia an intellectual high ground against those types of internet users who believe it is the obligation of content holders not to offend them (e.g. "You could have just hit the 'turn off bad images' button"), and could prove an excellent hedge against future demands for a "clean" (censored) Wikipedia. MalikCarr (talk) 03:02, 3 March 2008 (UTC)
If someone is willing to the script in other browsers (primarily IE6-7) and skins it could possibly be added as a Gadget and turned on through Preferences. (Gadgets also seem to load before site and user JS, so it would load sooner after the page loads) Though a software based preference would still be best. Mr.Z-man 06:03, 3 March 2008 (UTC)
Uh, that user scripts downloads a large non-cached page every time you open a page on Wikipedia. It's a nice idea, but in terms of performance, I do advise changing its design. GracenotesT § 16:25, 3 March 2008 (UTC)

This is not fair !!! I demand a random-badimage.js userscript to balance this emergence of selfcensorship ! :D rofl --TheDJ (talkcontribs) 11:34, 3 March 2008 (UTC)

I don't object to this development, but we must not allow the function of the badimage list to be altered by it. It should still be used solely to stem image vandalism - if we start to put Muhammad images or all nudity pictures on the list just so they can be hidden with this gadget, we're back to censorship, albeit opt-in. Happymelon 10:12, 4 March 2008 (UTC)
The point is well taken; a similar user script should be written for each different set of images, such as those of Muhammad, where a significant number of editors would like to opt out of seeing such images. (Happy-melon, you're not really saying that opting in to such a restriction is in some way censorship? Particularly if this was a gadget, there would be no way for other editors to know whether a particular editor has or has not chosen to avoid seeing a particular type of images.) -- John Broughton (♫♫) 15:19, 4 March 2008 (UTC)
What's wrong with opt-in censorship? Its one thing to tell people that "we won't censor things by default", its totally different to say that "you can't censor it just for yourself." Mr.Z-man
No, I'm not saying that someone indicating that they, personally, would like to take steps to avoid seeing an image, represents censorship. I'm saying that if we now take it upon ourselves to add all Muhammad images to the image blacklist, we are censoring their use on other pages. If someone wants to add Image:Maome.jpg to their userpage, as User:Mohummy has done (I just picked the first userpage which inlines that image, this argument is in no way affiliated to User:Mohummy), then WP:NOT#CENSORED clearly indicates that they can do so, and that no one has the right to stop them. If Image:Maome.jpg is on the image blacklist, it is not possible for the image to be used in that fashion, which is a clear violation of WP:NOT#CENSORED. Consequently, it's fine to have an opt-in script to disable all renderings of images which are already uncontroversially restricted, but adding new restrictions to otherwise unrestricted images just so they can be optionally disabled, is counter to our policies. I hope that makes my position clear. Happymelon 12:19, 7 March 2008 (UTC)

Current age in info box is out of date

I assume that the "birth date and age" tag in the info box of biographical articles is meant always to give the current age? Well it appears that it doesn't - I was just looking at the entry on Patsy Kensit and it says Born March 4, 1968 (age 39). Clearly it should have updated to "age 40" by now. 143.252.80.100 (talk) 22:14, 4 March 2008 (UTC)

Says 40 to me.↔NMajdantalk 22:28, 4 March 2008 (UTC)
This is probably because anonymous viewers may receive a cached copy of a page, even if the page contains templates which contain variables whose changes should have affected the content of the page. Page views aren't cached for logged-in users as the page content includes, among other things, links which depend on their user-name. Either way the "age" in the {{birth_date_and_age}} template will display properly for everyone after the next edit (on or after the subject's birthday) is submitted. — CharlotteWebb 14:55, 5 March 2008 (UTC)
I think it has to do with the day being (technically) different on the server, depending on where you are; for example, someone 12 hours away would be in a different timezone, thus in a different "day". The Wikipedia server is set to UTC, which is probably the cause. If you register, you can set your own time delay; there's one more benefit to having an account. ;) EVula // talk // // 05:55, 6 March 2008 (UTC)
Actually, the IP address above appears to be in or near London, UK, so local time and server time would probably be equal. — CharlotteWebb 16:07, 7 March 2008 (UTC)