Wikipedia:Village pump (technical)/Archive 17

When I try to use it it just pastes the entire wikitext (including noinclude and parameters) for a template instead of actually running the template and showing the resulting code. —Random832 15:22, 22 January 2008 (UTC)

I've noticed that substing templates haven't been working for a while there either (or through the API for that matter). I would have filed a bug, but every time I try to sign up for bugzilla I never get the email and I haven't been bothered to try emailing the appropriate people. Anomie 18:29, 22 January 2008 (UTC)
I've filed bug 12741; if you have an example of the API failing I can add that. —Random832 20:11, 22 January 2008 (UTC)
This and this should show the problem adequately. The template should be expanded, much as is done by the non-substed versions ([1] and [2]). That last link also shows the API having the same problem you describe with expanding templates. Anomie 20:55, 22 January 2008 (UTC)
Expandtemplates works now, your subst version still fails - file another bug? —Random832 16:33, 23 January 2008 (UTC)

Job queue age

Is there a way to see when the oldest entry in the job queue is from? —Random832 16:11, 23 January 2008 (UTC)

blockquote

I notice the <blockquote> tag works differently in Wikipedia than it does in html. According to Castro's HTML for the World Wide Web, blockquote should be used to enclose ordinary html markup; in particular, paragraphs tags should be used within blockquote tags.

If I try to indicate a paragraph within a blockquote in Wikipedia markup in the usual way, by leaving a blank line, it is ignored (see my sandbox. So what is the proper way to mark paragraphs within a blockquote? Where is the behavior of the Wikipedia blockquote tag documented? --Gerry Ashton (talk) 20:18, 23 January 2008 (UTC)

It's a bit ugly, but you can mark paragraphs with <p>, which still works in a blockquote. Algebraist 20:28, 23 January 2008 (UTC)
See bugzilla:6200. Sam Korn (smoddy) 20:51, 23 January 2008 (UTC)

What links here problem?

At [3], it looks like if an article, such as Joe Krol or Russ Jackson, transcludes a redirect to the template, they appear both as linking to the template AND as linking to the redirect. That's mildly annoying. --B (talk) 05:34, 24 January 2008 (UTC)

Text renders invisible when the line contains a "K"

I've been noticing strange behavior the past couple days on Wikipedia. In at first seemingly random patches or lines, text is rendering invisible (or white?). Then I noticed it seems to have something to do with the letter "k" in upper case (when I try to type it in upper case here, the entire line vanishes!). The text is there, because I can select it and paste it into another application to see it, but I don't see anything on the WP page or in edit mode. I'm running Firefox 2.0.0.11 on XP SP2, I haven't seen the problem using IE6, nor on any other website using Firefox. Here are some examples of where I encounter invisible text:

  • Under the above section "Suggestion for a new gadget", the entire second line of item 2 is invisible.
  • In that same section, item 9, 8th line: everything to the left of NF24's signature.
  • Under the section "Working WYSYWIG?" above, in the 3rd line, all the text to the left of "it's running", and in the next line, everything to the left of the signature "ais523".
  • Under Wikipedia:Translation/*/Lang/de#Geography, the 8th request in the list (Kaiserslautern), both instances of the word "Kaiserslautern" are invisible. This occurs for any entries on that page that contain a capital "k", and when I look at the page code (in edit mode), the same text is invisible there as well. (In this bullet, on the regular page, I cannot see any text to the right of the wikilink; in the edit window where I'm typing now, I can't see any text at all on the line that contains "kaiserslautern").

Anyone else running into this? Anyone have any ideas? I looked on Bugzilla and found nothing. -Eric talk 16:00, 21 January 2008 (UTC)

Update: The problem seems to be gone, maybe because I restarted my computer. I'd tried restarting Firefox before I posted here, with no effect. Sorry if this is a me-only issue. -Eric talk 17:06, 21 January 2008 (UTC)
The only thing I can think of is a stylesheet corrupted in a very interesting way. There probably is CSS somewhere for 'turn all text after a capital K in a tag white'; how it ended up being sent in the real stylesheet is another matter, though. I'm curious as to what the code is, now; I might try to find out. --ais523 15:23, 24 January 2008 (UTC)
Looking into it, there doesn't seem to be a selector that does exactly what you describe, so I'm confused again. --ais523 15:29, 24 January 2008 (UTC)
My only guess is a bizarre temporary glitch on my machine, if that's possible. I've never seen anything like it. At first I thought it had something to do with some possible experimentation on the main page code Jan. 20, because it started that day and seemed concurrent with some pages displaying with infinite width, i.e. showing a horizontal scrollbar that was impossible to bring all the way to the right. Haven't seen any more problems since Monday. -Eric talk 16:18, 24 January 2008 (UTC)

New preprocessor

The new preprocessor is now live on this wiki and http://test.wikipedia.org/ . Please read m:MNPP if you see any changes to article rendering. -- Tim Starling (talk) 10:54, 24 January 2008 (UTC)

Is this due to the new preprocessor? If not could someone try and fix it? Woody (talk) 11:51, 24 January 2008 (UTC)

This is misuse of {{!}}, one of the expected changes detailed in m:MNPP. They need to be fixed in the article or template where they occur. I can't fix all of them myself. -- Tim Starling (talk) 12:12, 24 January 2008 (UTC)
Tell me how and I will... Woody (talk) 12:21, 24 January 2008 (UTC)
Writing it out like that hasn't fixed all the other uses of Template:Asbox where they use the icon parameter. Perhaps the Asbox template will need to be changed so that the extra set of {{ and }} around the icon parameter are removed, then in all of its uses it can have the {{ and }} wrapped around the icon argument instead. (??) • Anakin (contribscomplaints) 13:35, 24 January 2008 (UTC)
Yes, I was wondering how to fix it. Each time I tried to use | | it didn't parse properly. When I then used the {{ }}, it worked but with an extra set of {{. According to Template:Asbox, the use of icon= is deprecated now anyway? (Am I reading that right). Woody (talk) 13:42, 24 January 2008 (UTC)
That's what I tried as well. The template's documentation doesn't suggest it's deprecated, but its example clearly gives the syntax that's now broken. I reckon the only thing to do is to remove the extra {{ and }} characters from the Asbox template itself and fix up any uses of the icon parameter in all the templates that use it. But I wouldn't want to mess about with it cluelessly - it could break goodness knows what. • Anakin (contribscomplaints) 13:58, 24 January 2008 (UTC)
You can't pass through bundled lists of parameters to split anymore, because with bug 5678 gone, there is no splitting mechanism. Like you say, the only thing to do is to change the way the template is invoked. This issue has been described on m:MNPP for over a week, and nobody has come up with a better idea. -- Tim Starling (talk) 14:15, 24 January 2008 (UTC)
I saw the
 | note      = This stub template is deprecated.
on the template documentation and presumed it was deprecated. On a random check, I couldn't find any others using the icon= parameter. I think we have to remove the forced {{ from the icon bit and then manually change them. I would be happy to go through them now, as long as someone knows how to edit the asbox template. Woody (talk) 14:18, 24 January 2008 (UTC)
There are only four or five that need updating (out of that group), on a quick run through with popups. • Anakin (contribscomplaints) 14:35, 24 January 2008 (UTC)

(unindent) Template:Asbox has now been updated, and the stub templates that were using it no longer rely on {{!}}. There are likely to other templates with very similar problems around though. • Anakin (contribscomplaints) 16:07, 24 January 2008 (UTC)

Page protection oddity with SVN currentversion of MediaWiki

Anyone noticed that protecting a page has two expiries?? Is this an error??
 

I noticed it when I used the latest version, 1.43.0-wmf.2 (ce9d259), as used here.

Thanks, --Solumeiras (talk) 17:09, 24 January 2008 (UTC)

This was an accident in rev:30110, reverted in rev:30121. Mr.Z-man 17:29, 24 January 2008 (UTC)

Slow images?

Is anyone else finding images loading very slowly at present, or is the problem at my end? – Tivedshambo (talk) 19:21, 24 January 2008 (UTC)

Same problem here. --TMF Let's Go Mets - Stats 19:32, 24 January 2008 (UTC)
Apparently most of the squid cache of images were lost during an update of the configuration; see [4] and [5]. —color probe (talk | contribs) 19:38, 24 January 2008 (UTC)

Semi-automated tagging of Shared IP Addresses

Betacommand has requested that I post for community discussion about a bot that I just proposed, called IPTaggerBot. If you are interested in commenting on the subject, please review the bot approval request at Wikipedia:Bots/Requests for approval/IPTaggerBot as well as the bot's userpage at User:IPTaggerBot. Thank you. Ioeth (talk contribs friendly) 20:40, 24 January 2008 (UTC)

Set of autoblocks on already blocked IPs

If a registered user vandalizes Wikipedia using an IP that is currently blocked with AO and ACB enabled (let's assume the account was created before the IP block was set or on another IP), and the account gets blocked itself now with ACB and autoblock enabled. Will this A) set an autoblock on that IP which would show up in the Special:Ipblocklist and B) disallow sleeper accounts to edit? As far as I know, write prohibition and block options are utilized in the following priority: account, IP address, IP range, autoblock, what would answer question B with 'no' and probably an autoblock isn't set by the software in such a case. Not sure if my imagination is correct. Thanks in advance --Oxymoron83 06:20, 25 January 2008 (UTC)

Customisation for the unregistered

At Talk:Main_Page#autofocus_the_search_box.2C_please an unregistered user was upset that there was no way for him to auto focus the search box (account holders can achieve this by editing their monobook.js).
There is currently no way for unregistered users to customise Wikipedia. It might be a good idea to make it customisable to some extent (auto focus; namespaces to search by default; thumbnail size; date auto formatting; font colour;) through cookies. This would be good for unregistered users as well as those who don't have javascript enabled. Any thoughts on this? Puchiko (Talk-email) 22:02, 12 January 2008 (UTC)

Given that it's almost trivial to register for a user account, and that IP addresses change, and that using a computer in a different location (library, work, home) will always involve different IP addresses, I really doubt that the developers are going to be willing to allocate any effort to this. I certainly wouldn't support it, given other priorities. -- John Broughton (♫♫) 16:08, 14 January 2008 (UTC)
Actually, cookieIP. The cookie stays for as long as you're using the same browser on the same computer, and as long as you don't delete it, or until it expires. It's what enables you to stay logged in.
You do have a point however, about there being more urgent tasks. Puchiko (Talk-email) 18:45, 14 January 2008 (UTC)
Or there might be a way to implement a little checkbox on the MP (or a little box that can be hidden, with the checkbox inside) that says "click here to make this page load with the focus in the searchbox (uses cookie)" and then leave a cookie on the reader's machine. I know several people who would like to do that without logging in. ~user:orngjce223 how am I typing? 02:55, 17 January 2008 (UTC)
Hear, hear! 68.101.123.219 (talk) 05:57, 21 January 2008 (UTC)
Apart from the main page search box, it could also be useful to allow unregistered users to set their text size, like many news sites do.
Anyways, if I want this implemented, what do I do now? Crosspost to WP:VPR? File a BugZilla request? Stick around waiting for some more discussion? Puchiko (Talk-email) 12:41, 25 January 2008 (UTC)

Watchlist

I know admins have a special unwatched pages list they can view. Lets say a person creates ana ccount, watchlists 500 pages, and then never logs in again. Would those pages never go back on the unwatched list? MBisanz talk 07:14, 20 January 2008 (UTC)

I don't know from experience, but logically the pages would remain off the unwatched pages list, since they are in the watchlist of some user. Even if the user doesn't log in ever again, the pages would still be on a watchlist, and therefore would logically be considered "watched". Of course, logic doesn't always work, but MediaWiki is usually logical (except for a few hard-coded namespace constants... ;-). Tuvok[T@lk/Improve] 07:29, 20 January 2008 (UTC)
Yep, that's the way it works. The unwatched page list simply tracks all pages that are not watchlist by any user. If one user is watching the page, regardless of that user's activity, it will not appear in the list. AmiDaniel (talk) 08:28, 20 January 2008 (UTC)
Would it maybe make sense then for the software to clear out the watchlist of any user who hasn't logged in in the last 180 days? Maybe even copy it to their user talk page in case they return. Is there an existing bugzilla on this or prior discussion? MBisanz talk 08:53, 20 January 2008 (UTC)

Special:Unwatchedpages is bascially useless anyway, because it only lists the first 1000 pages in alphabetical order (it's impossible to scroll any further through the list and get results). --ais523 14:57, 21 January 2008 (UTC)

It appears not to be working for the first 1000 and now just shows an empty list. Keith D (talk) 00:55, 26 January 2008 (UTC)


Is Special:Mostrevisions ever going to be updated? It's a very interesting page but could really use a new data dump (not updated in over a year). I searched for an alternative but all I found was Wikipedia:Most frequently edited articles, which is even more ancient. It would be nice if there was a script which could regenerate the data once in a while, but otherwise are there any equivalent stats anywhere else? • Anakin (contribscomplaints) 01:19, 24 January 2008 (UTC)

Well you can find out the number of revisions in any page by finding the diff between the first and the last edit - at the top there will be a number. Add 2 to that and you have the number of revisions in the page. Here's an example from this page. This won't work in rare cases where the revision has the wrong ID. This usually happens when a revision was deleted before MediaWiki 1.5 was installed on Wikipedia and undeleted later. The only article I know of to be effected by this is Air Force Amy - see this diff as an example. There are really no intermediate revisions by date at all. Graham87 02:00, 24 January 2008 (UTC)
Thanks, that's quite useful. I can use the data at Special:Mostrevisions, and get newer numbers from comparing revisions directly. Awkward but it would work. • Anakin (contribscomplaints) 13:29, 25 January 2008 (UTC)

Server error when editing PlayStation 3

I get an error when trying to edit PlayStation 3, since the new preprocessor is now live, I first guessed this was the cause, but Special:ParserDiffTest does not return anything suspicious. Anyone got an idea on how to fix it (and who we should bug). -- lucasbfr talk 11:27, 24 January 2008 (UTC)

I assume you mean 'Unrecognised punctuation character "$"' in the long page warning. It is related to the new preprocessor, I fixed it with this. -- Tim Starling (talk) 11:47, 24 January 2008 (UTC)
In fact, there is something worse: I am unable to save the page (Error: ERR_ZERO_SIZE_OBJECT, errno [No Error] at Thu, 24 Jan 2008 12:48:54 GMT). But thanks for fixing the long page warning. -- lucasbfr talk 12:50, 24 January 2008 (UTC)
Just to confirm that I am now able to save the page again. -- lucasbfr talk 14:05, 24 January 2008 (UTC)

This should hopefully be fixed now. We're back on the new preprocessor and I can save the page just fine. The underlying bug is in PHP, I've attempted to work around it. -- Tim Starling (talk) 08:44, 25 January 2008 (UTC)

MediaWiki:Ipbreason-dropdown

It seems the templated block reasons are now being expanded within the dropdown at Special:Blockip. I can't see any recent MediaWiki interface changes which would cause this. -- zzuuzz (talk) 12:30, 24 January 2008 (UTC)

I believe the new parser came online today. I've tried two attempts to make it appear as it did before, but it expands the template then inserts the wikitext that results, escaping entity references. I'm stumped as to what to do. Sam Korn (smoddy) 12:46, 24 January 2008 (UTC)
Fixed for the new preprocessor only. I switched us back over to the new preprocessor as soon as I changed the message. -- Tim Starling (talk) 08:39, 25 January 2008 (UTC)

Signature substitution issue

For some reason, this edit replaced a literal four tilde sequence in the source with my signature, though none of the preceding edits had done so. Can someone more MediaWiki-savvy than myself try to fix this? - (), 13:18, 24 January 2008 (UTC)

I guess that's due to the new parser. That's probably a bug, since the <includeonly> are not parsed inside a HTML comment. -- lucasbfr talk 13:43, 24 January 2008 (UTC)
Worryingly, it appears to be happening on all the other pages like that too: [6]. • Anakin (contribscomplaints) 13:45, 24 January 2008 (UTC)
I've reverted back to the old parser for now. -- Tim Starling (talk) 13:54, 24 January 2008 (UTC)
Old parser works: [7]. • Anakin (contribscomplaints) 14:03, 24 January 2008 (UTC)

Please consider fixing the 5678s, those will come back as soon as the parser is enabled again. I'm just worried about these new, unknown issues. -- Tim Starling (talk) 14:05, 24 January 2008 (UTC)

This is fixed now. -- Tim Starling (talk) 08:38, 25 January 2008 (UTC)

New preprocessor issue(s)

Ok, here is mine, but if you have broken pages, then you are welcome to add them here as well:

  • Special:ParserDiffTest/Portal:Current_events Current events has a problem. Seems to be caused by a change in behaviour on {{Start date}}. Special:ParserDiffTest/Template:Start_date There seems to be an extra newline. {{End date}} is also affected. --TheDJ (talkcontribs) 14:48, 24 January 2008 (UTC)
    • this would fix the template, but i'm just a tad worried on why the new parser thinks a new paragraph needs to be inserted here at all. It's just one new line, and not 2 newlines... --TheDJ (talkcontribs) 14:59, 24 January 2008 (UTC)
      • Report on meta --TheDJ (talkcontribs) 15:09, 24 January 2008 (UTC)
      • How often is onlyinclude used anyway? They were changing the includeonly/noinclude stuff; but I don't think I've ever seen onlyinclude in the wild before. I wouldn't be surprised if nobody had thought to test the new parser with respect to it... --ais523 15:16, 24 January 2008 (UTC)
        • Not only did I test the new parser with <onlyinclude>, I had to completely reimplement the feature, which was quite a bit more complicated in the new parser than it was in the old. I had to drop the capricious newline stripping, it wasn't practical to keep it. You'll have to put the </onlyinclude> on the same line as the end of the desired wikitext in the future, like how you do it already with noinclude/includeonly. -- Tim Starling (talk) 03:01, 25 January 2008 (UTC)
          • Thanks for the comments Tim ! --TheDJ (talkcontribs) 09:25, 25 January 2008 (UTC)

CategoryTree broken

C:CSD no longer renders in CategoryTree view. Instead, it renders as a mess of unparsed HTML. See the right column of my user page for how it looks like. I don't believe this is due to the new preprocessor... :) Pegasus «C¦ 09:31, 25 January 2008 (UTC)

Most likely the new parser that has gone online again over an hour ago. --TheDJ (talkcontribs) 10:00, 25 January 2008 (UTC)
Do you still have the problem? Everything seems fine to me atm. A purge must have taken care of it i think. --TheDJ (talkcontribs) 10:03, 25 January 2008 (UTC)
Still having the problem. :( Pegasus «C¦ 10:23, 25 January 2008 (UTC)
Should be fixed now. (r30148 synced 10:45) -- Tim Starling (talk) 10:48, 25 January 2008 (UTC)
Yes, it's fixed, and no purging was required. :D Pegasus «C¦ 10:49, 25 January 2008 (UTC)

Does anybody know how to create a footer with a lot of columns?

I was trying to create a template with 18 columns. Each column would only contain one letter so I think it wouldn't be too wide, but I don't seem to be able to figure out how to do this with the template coding. Remember (talk) 14:19, 25 January 2008 (UTC)

A B C D E F G H I J K L M N O P Q R S T

--MZMcBride (talk) 15:19, 25 January 2008 (UTC)

Please provide the necessary context when you ask a question. It appears from [8] that your question is not about templates, tables or columns in general. It is about one specific already existing template and giving a useful answer without knowing that is impossible. If you view the source of Template:Navbox with columns then you can see it contains code for each column and stops at col10. You can only get 18 columns in a use of that template if the template code is extended to 18 columns. PrimeHunter (talk) 19:00, 25 January 2008 (UTC)

extracting/mergine info from Wikipedia tables

I'd like to be able to manipulate info in tables in wiki pages, similarly to using a database system. For example, in the Comparison of file managers page, filter the Comparison_of_file_managers#Manager_views table according to the Comparison_of_file_managers#Operating_system_support table, such that, say, only linux file managers are displayed.

I know this is a lot to ask, but is there a technology available that would make this possible right from within Wikipedia? What's another simple way to do it? How about being able to extract the the info into a spreadsheet so I can do that manually? Sailor.nir (talk) 17:02, 25 January 2008 (UTC)

For a manual method, you can just copy and paste the contents of the table into Excel (if you have Windows & Excel) - it already knows how turn an HTML table into a spreadsheet. The contents of the tables aren't too amenable to filtering though, too many footnotes and stuff. Franamax (talk) 19:02, 25 January 2008 (UTC)

Page moves for watched pages

Is there any way to make it so that a page you are watching appears on your watchlist when it is moved. Quite often, I discover that a page I am watching has been moved, but I don't find out about it unless someone updates links to the page, or until someone subsequently edits the page. Thanks. DH85868993 (talk) 00:00, 26 January 2008 (UTC)

No. Unfortunately, page moves do not appear on watchlists. - FISDOF9 01:31, 26 January 2008 (UTC)

An odd username breaking the "user" template

User:-=CHAINSAW GRINGO=-'s user name doesn't work in the {{user}} template. When I type {{user|-=CHAINSAW GRINGO=-}}, it comes out as User-multi error: no username detected (help).. Corvus cornixtalk 00:40, 26 January 2008 (UTC)

If you use {{User|1=THE USERNAME}}, it should work. Let's give it a try, -=CHAINSAW GRINGO=- (talk · contribs). Ryan Postlethwaite 00:45, 26 January 2008 (UTC)
Yup, it worked - so just put 1= before it. Ryan Postlethwaite 00:46, 26 January 2008 (UTC)
Kewl, I never knew about that. Corvus cornixtalk 00:47, 26 January 2008 (UTC)
This is a general method to use template parameters containing the '=' sign without the software thinking '=' means parameter assignment. PrimeHunter (talk) 01:03, 26 January 2008 (UTC)

Template:Triple image seems to have broken, it doesn't like the px apparently. See Prairie Avenue for a good example. I can't see any changes using the Special:Parserdifftest. Is it related to the change? Woody (talk) 01:12, 26 January 2008 (UTC)

Never mind, user can't read template documentation. Knew it wasn't related to the change. Woody (talk) 01:16, 26 January 2008 (UTC)

Spam category

Strangely, Category:Happy Anniversary! was put on my user page (see my page), but I did not put it there. And it is not a template, or at lease one used often because I'm the only one to have it. What's going on? Who put me me in this useless category??? MAKE IT STOP!! —Coastergeekperson04's talk@12/29/2007 07:09

Resolved on IRC. Thanks weirdan7 and Dihydrogen Monoxide (I think, giggy). The problem was simply the UBX on my page. —Coastergeekperson04's talk@12/29/2007 07:37
Ready to archive. Pegasus «C¦ 11:19, 26 January 2008 (UTC)

Tor nodes

An ongoing discussion is in progress regarding adjusting the blocking policy in reference to TOR nodes. The discussion is here. Regards, M-ercury at 13:18, January 8, 2008

Ready to archive. Pegasus «C¦ 11:19, 26 January 2008 (UTC)

Ipblock exempt proposal

A proposal has started to allow established or trusted editors to edit via Tor, or other anon proxy. This discussion is located at

talk page

The proposed policy in its “needs to be worked on” form is located at

project page

Regards, M-ercury at 23:22, January 14, 2008

Ready to archive. Pegasus «C¦ 11:20, 26 January 2008 (UTC)

IE7 not displaying table properly

In Firefox (and I believe Opera), the rectangular title bar atop of this Infobox is rendered as expected: nice and tight with edges where they should be with no border, cellspacing nor cellpadding. It is just a simple table of three rows and two columns, some cells spanning others. However, when I looked at how it appeared in IE7, it having caused me layout frustrations before, it was no real surprise to see it could not render it correctly. I edited the template in IE using <tr>s and <td>s, along with explicit height measurements, and previews showed no change at all. I also tested it with line-height:0px (and 1px in case it didn't recognise 0px) styling, since it appears to be copying the text height (or close to it), but still no change. Being such a simple table, I would have thought that IE7 would be able to do this without a drama, but no. Can anyone see a fix for this? I thank you in advance (again).  SEO75 [talk] 19:52, 24 January 2008 (UTC)

Wikitables and images don't mix very well. I'll see what I can do. EdokterTalk 15:54, 25 January 2008 (UTC)
Best I could do... I simplified the table to use styled border instead of images. It still looks the same in Firefox, but get a 1px gap on top of the image in IE, which I can't seem to get rid of. But it is an improvement none-the-less. EdokterTalk 23:12, 25 January 2008 (UTC)
*slaps self* Why didn't I think of that?.. Thank you very much! :)  SEO75 [talk] 15:27, 26 January 2008 (UTC)

Another questions

I just created this compact template below. Anyone know how to make the default version of this hide instead of show?Remember (talk) 20:07, 25 January 2008 (UTC)

Try using the template {{hidden}}. AmiDaniel (talk) 20:32, 25 January 2008 (UTC)
This seems to do it. Anomie 20:33, 25 January 2008 (UTC)
Imho it was a terrible idea to use style="display:none" in navigation boxes: users with JavaScript disabled will not be able to see the content at all. Until this is fixed, it's better to use collapsible tables with extra collapsed class ∴ AlexSm 21:39, 25 January 2008 (UTC)
That's the way the NavFrame code in MediaWiki:common.js works, look at line 275. Anomie 02:49, 26 January 2008 (UTC)
So? That's exactly my point: this was a bad thing to do. I did it properly in ru:MediaWiki:Common.js by introducing collapsed class for NavFrame (just like for collapsible tables), but nobody was interested in fixing it in English Eikipedia: see MediaWiki talk:Common.js/Archive Nov 2007#CSS hidden NavFrameAlexSm 03:21, 26 January 2008 (UTC)

I'd support you Alex... If you would ever consider proposing it again. I think you are right and your technique should be implemented. --TheDJ (talkcontribs) 01:53, 27 January 2008 (UTC)

GreaseMonkey focus script

I've created a script that Firefox users with Greasemonkey can install, so that Wikipedia:Main Page FAQ#Why doesn't the cursor appear in the search box, like with Google? can suggest an alternative for users who do not want to create an account but happen to have the extension. The script is located at Wikipedia:Main Page/searchfocus.js, and the link to follow to install it is (sans line break):

http://en.wikipedia.org/w/index.php?title=Wikipedia:Main_
Page/searchfocus.js&action=raw&gmname=searchfocus.user.js

(clickable) The fake parameter at the end causes Greasemonkey to prompt to install the script when the link is clicked. For security reasons, the page should be protected. Does anyone have any suggestions before it goes "live"? Is it a bad idea? Should the script be in the MediaWiki namespace? GracenotesT § 01:30, 26 January 2008 (UTC)

Okay, I've added the link. Please comment if you have any suggestions whatsoever ^_^ GracenotesT § 01:12, 27 January 2008 (UTC)

According to this site, the site is blocked by [9]. However, I don't see it listed anywhere. If it is listed, I was wondering if the site could be removed, as I would like to use it to test whether templates work in different browsers. Thanks. SharkD (talk) 07:36, 26 January 2008 (UTC)

Oh, lol, I just posted a duplicate of this below, but it's a little more detailed as to how to fix the problem. :P Cheers =) --slakrtalk / 20:19, 26 January 2008 (UTC)

References and notes

Is there any way in which to have separate references and notes sections within an article, without resorting to cumbersome hand coding? Specifically I'm editing this article: List of West Midlands railway stations. TicketMan - Talk - contribs 18:07, 26 January 2008 (UTC)

You could use {{ref}} and {{note}} for the alphabetical notes. –Pomte 18:08, 26 January 2008 (UTC)
Fantastic - many thanks for the speedy reply, Pomte. TicketMan - Talk - contribs 18:16, 26 January 2008 (UTC)

Flags messing up articles using Template:Infobox Wrestler

Flags seem to be messing up articles using Template:Infobox Wrestler. For instance, with the Japanese flag it displays like this for some reason. This is happening for many flags. Removing the flag will get it back proper for instance like this, but this only started occuring last night. None of the templates have been changed, so I'm struggling to see what the problem is. Please also see this discussion at WT:PW. Cheers, D.M.N. (talk) 20:55, 26 January 2008 (UTC)

The problem seemingly have been fixed. D.M.N. (talk) 21:45, 26 January 2008 (UTC)
The problem was that Template:Infobox Wrestler had a line of html in the <table> header:
summary = "Wrestler's personal file: {{{name}}}"
I had updated the flag templates yesterday to wrap the image rendering in a CSS class, so when a {{flagicon}} was part of the {{{name}}} parameter in this template, it expanded with a class="flagicon" as part of the flagicon markup, and the quote character fouled up the summary line. I'm not sure why the summary markup was actually needed, so I just commented it out. — Andrwsc (talk · contribs) 22:19, 26 January 2008 (UTC)

Need help with JavaScript

  Resolved

I could use some help with JavaScript here. I've been trying to set up a system where I can enable and disable individual scripts without making an edit each time. To accomplish this, I set up a local JavaScript file with a variable for each script I use, and then made each script in my monobook.js page check that variable for whether it should run or not.

Links:

The problem is, when I saved the page and bypassed my browser cache after setting up the scripts to check the local variables, my entire monobook.js page stopped working; something caused all the JavaScript to break. I checked very carefully that the local JavaScript file was indeed running, so it must be a problem with my second edit that had the scripts check that local file.

Here's the source of the local file:
var local_all_scripts    =     true;    // turn all scripts on or off

var local_navpops        =     true;    // Lupin's navigation popups
var local_shortcuts      =     true;    // User:Jsimlo/shortcuts.js - customized sidebar links
var local_user_watchlist =     true;    // User watchlist from [[User:Tra/userwatchlist.js]]
var local_logs_link      =     true;    // Logs link in toolbox
var local_compare_link   =     true;    // Reformats compare button in history as a link
var local_formatting     =     true;    // Performs basic formatting changes from a tab while editing
var local_metadata       =     true;    // Metadata assessment script
var local_sixtabs        =     true;    // Adds extra edit/hist tabs for both discussion page and article page (6 tabs)
var local_history_format =     true;    // Improves formatting of history pages
var local_searchbox      =     true;    // [[User:GeorgeMoney]] script improves search box functionality
var local_edittop        =     true;    // Edit top link by [[User:Alex Smotrov]]
var local_unwatch        =     true;    // Unwatch links next to page names in watchlist from [[User:Alex Smotrov/wlunwatch.js]]
var local_S_tags         =     true;    // User:S tagging script
var local_newpagebox     =     false;    // Box with newest pages in sidebar, disabled by default
var local_recentbox      =     false;    // Box with recent changes in sidebar, disabled by default
var local_logtable       =     true;    // Nice table formatting for [[Special:Log]]
var local_twinkle        =     true;    // Twinkle script (reverting, warning, reporting, csd, xfd, prod, rpp)
var local_tw_rollback    =     true;    // Rollback integration with Twinkle
var local_wikEd          =     true;    // [[User:Cacycle/wikEd]] in-browser text editor
var local_friendly       =     true;    // [[WP:FRIENDLY]] script similar to Twinkle
var local_VoA_helper     =     true;    // [[User:Voice of All]]'s helper JavaScript
var local_UTC_clock      =     true;    // UTC time clock with double-click purge function
var local_VoA_history    =     true;    // Some history-related script from VoA
var local_Lupin          =     true;    // [[User:Lupin/recent2.js]] anti-vandalism script
var local_changelinks    =     true;    // Modified version of [[User:Voyagerfan5761/changelinks.js]]
var local_GeorgeMoney    =     false;    //[[User:GeorgeMoney]]'s helper scripts (Cookies, AddLinks, and Include)
var local_rollback       =     true;    // Rollback with summary
var local_scriptcontrol  =     false;    // ScriptControl, currently in early stages of writing it

As you can see, the local file is just a series of variables, so it can't be the problem. Could someone more experienced with JavaScript explain why this would cause my scripts to break and how to get this to work? Pyrospirit (talk · contribs) 18:25, 25 January 2008 (UTC)

  • "Something caused all the JavaScript to break" is not good enough: get Opera or Firefox and then use "Error console" to track down the exact error message.
  • Insted of using variables, you can simply move all importScript statements into this local file, and then comment/uncomment them.
  • If you do use variables, your code should also work when variables are not defined (e.g. local script file is missing): use something like if (window.local_navpops) ...
AlexSm 21:36, 25 January 2008 (UTC)
I'll check the error console in a minute. I'd prefer to use variables because that way, I still have the scripts when I use a different computer. What I do in case the variables are undefined is this: if I want it enabled when the variable is undefined, I run it if the variable !=false, and if I want it disabled when the variable is undefined, I check for ==true instead. Since it's neither true nor false when undefined, this should give a default setting for use on other computers. Pyrospirit (talk · contribs) 21:57, 25 January 2008 (UTC)
I just checked the error console, the error was "local_all_scripts is not defined". But that's the first variable I set in the local file, so it certainly seems to be defined. What's up with that? Pyrospirit (talk · contribs) 22:04, 25 January 2008 (UTC)
It seems like there is no consistency whether the browser waits for a new linked script to load (like Opera 9) or simply continues with the next statement (like Firefox 2) ∴ AlexSm 01:56, 26 January 2008 (UTC)
Using undefined variable triggers a JavaScript error. That's why you either check for (typeof your_var == 'undefined') or simplyt use window.your_var : every global variable is also a property of a window object, and empty properties return null without errors ∴ AlexSm 01:56, 26 January 2008 (UTC)
Thanks for the help with that. I've converted the variables into properties of the window object as you suggested, and my scripts are no longer broken. However, editing the local file and saving it appears to have no effect on my scripts; it's as though it is reading all the variables set in the local file as null instead of the proper boolean value I assign them. Any idea why? Pyrospirit (talk · contribs) 02:18, 26 January 2008 (UTC)
Because you are using document.write() to set the local_ variables, the local script will be injected into the input stream after the current (monobook) script finishes executing. You need to either put the local_ variable settings into monobook or else move all the importScripts into a file that gets executed after your local variable script, so try creating a User:Pyrospirit/setscript.js file and in monobook.js add a document.write() just after localhost/scripts.js to run pyrospirit/setscript.js. It's the execution order that's killing you. Put some alert()'s in so you can check the order everything is happening.
You could also try using importScript("h-t-t-p://localhost/scripts.js") instead of document.write(), that might get scripts.js executed in the right place (no idea if that will work, throw in some more alert()'s). Franamax (talk) 02:44, 26 January 2008 (UTC)
I already stated this above: only Opera 9 seems to wait for a new external script to load and execute, while Firefox and IE6 continue the current script and then the localhost/scripts.js is simply executed too late ∴ AlexSm 03:29, 26 January 2008 (UTC)
Yes you did say that Alex, I was just trying to give Pyro some more hints on tracing the execution sequence and watching when variables get instanced. FWIW Opera 9 should maybe not immediately load and execute a script called in a document.write(), logically it should be written at the end of the incoming byte stream, shouldn't it?
Would you agree that my advice to Pyro was roughly correct as far as changing the execution locations for testing logical variables, and using alert()'s for tracing? I'm certainly not a .js expert, I'm just trying to push across some very well-worn experience, and I'm always very happy to learn more. :) Franamax (talk) 06:47, 26 January 2008 (UTC)
Bit of a tangent but: it often is advantageous to wait for document load before execution of anything besides scriptloads or variables (via addOnloadHook), although even this is not 100% -> bugzilla:12773. --Splarka (rant) 08:03, 26 January 2008 (UTC)
O-w-www my brain hurts :) Splarka, for Pyro's benefit (maybe), do you have any code-specific suggestions as to how to "wait for document load" before executing another script? More suggestions for how Pyro can accomplish the desired selective-enabling of .js functions? I'm interested in the answers also.
And I would be most pleased to follow up with you at one or other user talk page on pretty much everything you just said so I could understand just what the hell you're getting at. :) Not understanding a word has always been my gateway to knowledge, but it's probably not appropriate to this thread. I will chase you later. Franamax (talk) 08:39, 26 January 2008 (UTC)
I did the thing with alerts as you suggested, and apparently by the time the local script variables start loading, all the other scripts have finished loading, making the variables meaningless. So, all we need to do is figure out a way to make the local variables load before anything else even starts loading, and then it should all work. I'll try the importScript suggestion, but if that doesn't work, the solution might be more complicated. Pyrospirit (talk · contribs) 16:35, 26 January 2008 (UTC)
I tried switching to the same code used by the importScript function, but that didn't work either. It's still loading the local variables after the rest of the scripts are finished. Anyone know how to fix this so that the local variables load before anything else? Pyrospirit (talk · contribs) 01:23, 27 January 2008 (UTC)
Did you try my other suggestion, using two document.write()'s in a row, one to execute local/scripts.js and the second to execute pyrospirit/setscript.js? If you move all the importScript code from monobook to setscript.js, that should execute in the right sequence. Franamax (talk) 02:17, 27 January 2008 (UTC)
It seems to be working as it should now. Thanks everyone for helping me get this right! Pyrospirit (talk · contribs) 04:41, 27 January 2008 (UTC)
I'm not sure that's the most elegant solution, but hey, if it gets the job done! My quibbles would be using the same name for a script on two different platforms, shouldn't one be called "local_script.js" and the other be called "server_script.js" or words to that effect? And should you be relying on a local httpd server running, or instead using a "file:///C:/..."-type address to get the local script directly from the file system (your choice there)? Anyway, thanks for the directions to all the useful scripts spread over wiki and now I can watch your .js pages to learn the latest cool stuff, helping works both ways! Franamax (talk) 11:56, 27 January 2008 (UTC)

Move page "links" function broken

When one has successfully moved a page, one now (since the preprocessor upgrade) gets a message like this:

The page "old blah" (links) has been moved to "new blah" (links, revert).

Please check whether...

Note the destinations of several of the links which appear to be untranslated programming code. Is this a known issue? Orderinchaos 07:03, 26 January 2008 (UTC)

Ilmari Karonen fixed this a bit ago. --MZMcBride (talk) 00:23, 28 January 2008 (UTC)

Monobook.js not loading

I have no idea why my monobook.js is not loading since this afternoon. Firefox refuses to load it even after repeatedly Shift-Ctrl-Reloading any Wikipedia page.

System configuration is Ubuntu Gutsy 7.10, with Firefox 2.0.0.11. I have NoScript installed but it's set to allow scripts for wikipedia.org and wikimedia.org. Lastly, when I view source of this page, I see the following which should load monobook, but now doesn't.

<script type="text/javascript" src="/w/index.php?title=User:Pegasus/monobook.js&action=raw&ctype=text/javascript&dontcountme=s"></script>

(Note that the &amp; thing is escaped in the original page source so you've to see my comment in the edit window to see what the line actually is.) Pegasus «C¦ 11:10, 26 January 2008 (UTC)

It's likely your browser is, or our squids are, caching the result; so, you'll need to navigate directly to that URL and hard refresh the page. --slakrtalk / 20:17, 26 January 2008 (UTC)
I have the exact same problem (and mostly same setup: ubuntu 7.04, firefox 2.0.0.11). Nothing in my monobook.js file is working (tried multiple hard-refreshes/cache-clears). Any other suggestions? -- Quiddity (talk) 21:19, 26 January 2008 (UTC)
Most browsers have error consoles .... hint hint. :D --TheDJ (talkcontribs) 01:58, 27 January 2008 (UTC)
Now the problem is gone. And perhaps not unexpectedly, there's nothing untoward logged in the error console. Pegasus «C¦ 09:44, 27 January 2008 (UTC)
Ditto. Sorry I couldn't diagnose it before it went away.. -- Quiddity (talk) 19:33, 27 January 2008 (UTC)