Jump to content

Wikipedia:Village pump (technical)/Archive 15

From Wikipedia, the free encyclopedia

Getting the [hide] and [show] buttons.

I'm having trouble finding the markup to display the [hide] and [show] buttons, like on Template:Navbox and Template:Tools. I have looked at the source for those particular templates, but I still cannot find the code. Does anyone know it? Thanks. Also, if you could reply here and on my talk page, that would be great, because I have a liablilty to completely forget where I ask things. Thanks again. Asenine (talk)(contribs) 13:04, 10 January 2008 (UTC)

Please see the MediaWiki:Common.css NavFrame/NavContent/Navbox etc CSS classes and several Javascript classes in MediaWiki:Common.js. These are required to make the these systems work. --TheDJ (talkcontribs) 13:38, 10 January 2008 (UTC)
See Wikipedia:NavFrame and Wikipedia:Collapsible tablesAlexSm 15:26, 10 January 2008 (UTC)

RESOLVED: Watchlist truncation

 Works for me

Anybody know why my watchlist suddenly only shows 14 changes, all on today? I normally see three days worth. I saw plenty of changes yesterday, so I'm not just watching idle pages. I'm logged in. I tried clicking the "7 days" link. The URL now has "days=7" in it. Tried clicking the "Show bot edits" link. Other edits are not hidden. I've tried doing a forced-reload of the page (SHIFT+CTRL+R). Tried restarting my browser and clearing my cache. Nothing helped. Page says "last 250 changes, as of 13:17, 10 January 2008" and "617 pages on your watchlist". Namespace is "all". Screenshot; copy of page HTML. Firefox 2 on Linux. I'll try another PC and netfeed when I get into work in a bit, but this seems sufficiently weird that I wanted to report it. —DragonHawk (talk|hist) 13:33, 10 January 2008 (UTC)

In your account preferences you might try to disable: "Enhanced recent changes (JavaScript)". Perhaps something is wrong with that option causing the HTML to break ? --TheDJ (talkcontribs) 13:40, 10 January 2008 (UTC)
Never mind. I realize now that there have been a truly insane amount of edits on one of the pages I was watching. Disabling the JavaScript-collapse revealed that. Removing the hyperactive page from my watchlist fixed it. Nothing to see here, move along.  :) Thanks! —DragonHawk (talk|hist) 13:51, 10 January 2008 (UTC)

I'm trying to write a script to provide a JavaScript prompt box for an edit summary when using the rollback feature. I'm wondering if anybody has any tips for getting at the rollback links on the history, diff, and user contrib pages. What I have so far is at User:Voyagerfan5761/rollbacksummary.js. Complex DOM scripting is above what I've attempted so far, so I need a little nudge in the right direction. Anything will be much appreciated. I'll be offline for about an hour, then back to check on an answer. Thanks in advance! Tuvok[T@lk/Improve] 23:58, 10 January 2008 (UTC)

I thought rollback was a direct query you cannot control (as in, you can't give an edit summary at all, it is automatically generated by the wiki)... -- ReyBrujo (talk) 00:20, 11 January 2008 (UTC)
@ReyBrujo: No, you can specify a summary by appending "&summary=foobar" to the URL or POST parameters. @VoyagerFan5671: You look to be more or less on the right track with your script there, and I'm not sure exactly what it is you're having difficulty with. Can you be a tad more specific? AmiDaniel (talk) 00:34, 11 January 2008 (UTC)
Good to see you around, AmiDaniel! What I'm really wondering is whether there's an easier way to do things than the way it is now. It all seems like such a kludge at the moment. I'm continuing work on it; I hope to be able to test it on one of my subpages by the end of the night. Without errors. ;-) Tuvok[T@lk/Improve] 01:01, 11 January 2008 (UTC)
Actually, I have a specific question now. I'm trying to insert a modified copy of the rollback link after the one on the history page, but for some reason it loops and I have to stop the script. I end up with hundreds of identical links. What am I doing wrong? Is it something to do with appendChild( newNode ) or what? Tuvok[T@lk/Improve] 03:09, 11 January 2008 (UTC)
You are doing a for cycle up to links.length, but add a new child whenever you find the action=rollback string, so links.length increases, and you continue adding children to links.
And AmiDaniel, thanks for the tip, I will check that out later. -- ReyBrujo (talk) 03:35, 11 January 2008 (UTC)
Hmm, it seems links shouldn't be automatically slurping up the new rollback links. I hadn't thought of that, since it's rather illogical for a variable to change without being assigned a value (I'm not aware of any assign-by-reference capability in JavaScript). Looks like I'll have to change the URL slightly and add a negation to the regex so as not to add a link after it's already been inserted. Thanks! Tuvok[T@lk/Improve] 06:21, 11 January 2008 (UTC)
Hope you don't mind that I did a little meddling with your script :) Fixed some odd behaviors with the click events and with loading, but didn't run across the infinite loop problem (except when I accidentally introduced it myself :D). The act of locating the links can be done far more elegantly and efficiently using the className as opposed to the parameters in the URL. That is, all rollback links are in the class "mw-rollback-link", and only rollback links are in that class. Additionally, your script really should add the links outside of that enclosing span, to ensure that this rule holds true. I took a whack at it here, but it's four am, and I've had a few glasses of wine :) You get the idea though. Otherwise, works as specified. AmiDaniel (talk) 11:12, 11 January 2008 (UTC)
Note also the built-in function getElementsByClassName (not a function of the document object) from wikibits.js. It's useful for code-minimizing, since not everything useful in MediaWiki has an id (and some useless things do). GracenotesT § 16:39, 11 January 2008 (UTC)
I saw that in my watchlist, AmiDaniel; I don't mind at all. :) I've asked for help from Ioeth on getting the links to go in better places (using XPath, probably), so we'll see how that goes; I'm on my way to check that page after I'm done here. (The links aren't inserted quite the way I'd like, either.) And it looks like I neglected to post here again after I'd solved the looping problem. That's why you didn't get it until you changed the code yourself; I'd come up with a solution (perhaps a slow one; see below) already. Many thanks for adding the return statements, though. :-) Tuvok[T@lk/Improve] 21:02, 11 January 2008 (UTC)
One nice trick that avoids the infinite loop problem without extra tests is to run your loop "backwards": for (var linkid = links.length - 1; linkid >= 0; linkid--) This way, when you insert a new link after the current one it doesn't matter that it is inserted in links right away because you're already past that point. As a bonus, it may be faster (maybe even noticeably so) since it doesn't have to re-evaluate links.length on every iteration.
As for why links grabs the new links immediately, the return value of getElementsByTagName isn't actually an array; it's a NodeList, which is defined to be "live" in exactly that manner. Anomie 18:04, 11 January 2008 (UTC)
Ah, thanks Anomie! Good explanation, and a neat little trick. I might put that into the script at some point, and get rid of the &js=y part. That probably is slowing me down with more regexes than are strictly necessary. I knew I could come here to ask questions; en.wp is the best place I've found for JavaScript help. Must be all those userscripts... ;-) Tuvok[T@lk/Improve] 21:02, 11 January 2008 (UTC)
Just an update: I've implemented Anomie's suggested backward loop and taken out a couple of the regex tests. It does indeed work noticeably faster. Thanks again! Tuvok[T@lk/Improve] 21:08, 11 January 2008 (UTC)

500 Internal Server Error

While doing some editing on a new Dynamo article, when I submitted my edits I would get a 500 Internet Server Error when I submitted the page. This happened just now so if someone wants to take a look in the server logs.. (2:00 AM central time, USA)

I tried opening a new edit window, copied and pasted what I had done, and saved it. It worked. Then I realized I replaced the entire article with one section, so I edited the previous version, copied in the new edit in the correct spot, added a comment with a comma, hit save.... and another 500 Internet Server error.

I took the comma out of my comment and it saved. Is there problem with commas in comments? I tried adding a comma'd comment in the Sandbox, and it worked.

Overall, I don't have the slightest idea what the problem is. Someone with webserver log access needs to check this. DMahalko (talk) 08:09, 11 January 2008 (UTC)

That's a rather common occurrence (far less common these days!), and really isn't cause for concern. It's undoubtedly completely unrelated to there being commas in comments; rather, it happens when servers crash while you're in the middle of saving a page, etc. AmiDaniel (talk) 11:15, 11 January 2008 (UTC)

Cannot disable Navigation popups

Hello, A (very long) while ago, I edited my monobook.js file so that I could use navigation popups, using a script another user had written. However, I've never really liked the popups because they usually tend to cause my browser (Opera 9.23) to stall for a bit when loading and always ended up blocking parts of the article I was actually trying to read whenever I accidentally moused over a link. More recently they've started popping up while editing pages and I am actually prevented from editing the page anytime one pops up which is incredibly annoying. They take forever to go load anyway so they're pretty useless, plus they take forever to go away so if I want to read what's underneath it or continue to edit a page I'm SOL for a while. So I've finally decided to turn them off. I removed the script from my monobook.js file but to my dismay, I kept getting the popups. I checked my preferences and in the "gadgets" tab there was a checkbox for "use navigation popups" but it was unchecked so I don't see any way they could or would still be on. What is causing this and how do I stop it? Thank you for your help. Uniqueuponhim (talk) 01:39, 6 January 2008 (UTC)

Try following the steps listed here. Tra (Talk) 01:45, 6 January 2008 (UTC)
Thank you. Uniqueuponhim (talk) 02:35, 13 January 2008 (UTC)

Rollback for non administrators is now live. If you feel you need the tool, and can use it constructively, please file a request at the above page. Ryan Postlethwaite 23:30, 9 January 2008 (UTC)

Before doing so, you may wish to consider the legality of the decision to implement this system following a discussion at which the support-oppose ratio was only 2:1. Discussion is currently taking place at Wikipedia talk:Non-administrator rollback. — Hex (❝?!❞) 01:24, 10 January 2008 (UTC)
2:1 is enough consensus, please, just get on with things now - if there's problems, it will be stopped. Ryan Postlethwaite 01:28, 10 January 2008 (UTC)
I believed you were being honest with things when Doc accused you of being sneaky with the poll. Now I'm not so sure. -- Ned Scott 02:02, 10 January 2008 (UTC)
Sneaky with the poll? What weight do you honestly think I have with the developers? None whatsoever. I didn't even close the discussion, an opposer did. Ryan Postlethwaite 02:31, 10 January 2008 (UTC)
I'm not saying you slept with anyone or paid them off with a suitcase full of cash, but you're just so quick to dismiss the concerns about what happened here. I guess I shouldn't take it out on you, though. -- Ned Scott 02:41, 10 January 2008 (UTC)
I was only acting on people quickly trying to dismiss the proposal, when on the basis of the pole, the majority of users wanted the tool. I actually don't care too much about it, if someone really wants to stop it, then go ahead - I've got rollback, it doesn't mean much to me. Ryan Postlethwaite 02:44, 10 January 2008 (UTC)
Legality?? Cripes, this is a wiki; we try things, evaluate them, and either improve or rollback. This page is clearly not another avenue for discussing the consensus. Ryan has *notified* this page for *technical* discussion. John Vandenberg (talk) 02:47, 10 January 2008 (UTC)
EVERYBODY CALM DOWN. ok, are you calm now? listen to this: The developers did not implement the proposal that was voted on. They implemented a single technical feature. There was a wide consensus, including many of the oppose voters, that we should have some form of this feature. If you don't like the proposal, you're free to suggest alternatives for what the policy or process or whatever should be for giving rollback. We are not bound to the proposal. —Random832 02:53, 10 January 2008 (UTC)
for reference. Cheers =) --slakrtalk / 05:41, 10 January 2008 (UTC)

Bug?

Wikipedia:Requests_for_rollback#User:j.sweeton.40wnri.com - It seems that because of this user's username (?) admins are getting a permission error on trying to grant rollback. Pedro :  Chat  13:27, 10 January 2008 (UTC)

I've asked on meta.Random832 15:08, 10 January 2008 (UTC)
Sorry, for stewards rollbacker-usergroup is not available from meta, best regards, --birdy (:> )=| 15:21, 10 January 2008 (UTC)
Reported as bug 12602. Just a note to stewards: Grant yourself steward rights on enwiki, then change the rights for the user from here. Prodego talk 18:54, 12 January 2008 (UTC)
Hasn't worked either. —DerHexer (Talk) 20:43, 12 January 2008 (UTC)

Unviewable, uneditable "Open tasks" section on tk pg w/ nested templates re "scope of the following WikiProjects"

_ _ Talk:Edward Carson, Baron Carson, as seen by those who ask the server to number the headings, skipped section 1 and its subsections. It also had no ToC in spite of 7 visible headings, presumably for all users.
_ _ I added a __TOC__ directive just above the first editable/visible section following the lead secn. This did its job of forcing the TOC to display, revealing an "Open tasks" secn (presumably from a template invoked within the nest of templates that were the whole lead secn and are now accompanied by my TOC directive) with 9 subsections. However, those ten headings (and any text within their sections) are still invisible, and undetectable except by their ToC entries and the odd numbering of the other hdgs and ToC entries.
_ _ Moreover, re the anomalous sections

  1. clicking on the ToC entries for them produces no discernible effect
  2. they are neither viewable nor editable by editing the URL for the page, since "...edit&section=1" turns out to refer in this case to the first visible section following the lead 'graph.

--Jerzyt 20:44, 11 January 2008 (UTC)

The problem is that {{WikiProject Unionism}} includes Wikipedia:WikiProject Unionism/Tasks, and that page contains several headers. The fix is to remove the headers from that page. Anomie 21:12, 11 January 2008 (UTC)
I've notified WikiProject Unionism about their issue. Anomie 21:25, 11 January 2008 (UTC)
(edit conflict) Talk:Edward Carson, Baron Carson transcludes {{WikiProject Unionism}} which transcludes Wikipedia:WikiProject Unionism/Tasks which contains <includeonly>__NOTOC__</includeonly>. It was added in [1]. I don't know why. PrimeHunter (talk) 21:17, 11 January 2008 (UTC)
I don't want to get my hands even dirty enuf to fully grasp Anomie's explanation at Wikipedia talk:WikiProject Unionism#Your "Tasks" page is breaking talk pages, and so thanks to everyone who is willing to dive in. But i have put a harm-reducing note on the Campbell-bio talk page.
--Jerzyt 02:02, 12 January 2008 (UTC)
Fixed per request. Anomie 14:10, 12 January 2008 (UTC)

Possible to transclude only part of a template or page?

Is there a means within mediawiki to transclude only a portion of a template or article into another article? For instance, if my template or article was composed of a table, is there a means to selectively transclude portions of that table into another article? Cadrm1n (talk) 02:05, 12 January 2008 (UTC)

Yes. Use <noinclude></noinclude> tags on the parts you don't want to transclude. - FISDOF9 02:09, 12 January 2008 (UTC)
Thanks for the quick response, but not sure that would work for what I have in mind; in the example, let's say I have a template {{TemplateABC}} containing a table, 3Rx3C, and I ultimately want to parcel out the information in that table into multiple other articles. For instance, In article A, I want to bring in all of ROW 1 from that table. In article B, I want to bring in all of Row 3. In article C, I want to incorporate all of COLUMN 3 from that same table. In summary, I'm looking to keep some information all updated in the table, and then parcel out different portions of that information to multiple external articles. How would I call that template from within articles A, B, and C, and specify the desired subsets of information? I can't very well incorporate <noinclude></noinclude> tags within the traditional {{TemplateABC}} template call. If you can point me to an example of what you describe in action, it would help me understand how that approach may or may not work. Cadrm1n (talk) 03:13, 12 January 2008 (UTC)
The only way I could see it working with existing wiki-syntax would be to have {{TemplateABC}} itself transclude subtemplates for each cell, say {{TemplateABC/R1C1}} thru {{TemplateABC/R3C3}} and then have articles A, B, and C then transclude only the cells of the rows and columns pertaining to that article. Caerwine Caer’s whines 03:24, 12 January 2008 (UTC)
There is an extension to transclude marked-out sections of articles, but it's not installed here. Tuvok[T@lk/Improve] 03:54, 12 January 2008 (UTC)

You can also use <includeonly></includeonly> as an alternative to marking off the unwanted parts. -- Ned Scott 05:55, 14 January 2008 (UTC)

Wikipedia crawling

I need to retrieve several hundred Wikipedia articles, and I would like to do so via the Special:Export function. Is there any problem with doing this? Does Wikipedia have a preferred way to do this that would be better? —Preceding unsigned comment added by 66.142.230.228 (talk) 04:06, 12 January 2008 (UTC)

No it's not a problem. Note that Special:Export can export multiple pages per request. It's probably best to do the job in batches of about 20. -- Tim Starling (talk) 08:34, 12 January 2008 (UTC)
The MediaWiki API can also be used to do this in a more diverse array of formatting (JSON, XML, PHP, etc.), if you add "titles=Page_1|Page_2|[...]|Page_19|Page_20" as a parameter. It works by GET or POST. Special:Export is just as good, but that's mainly to be used with Special:Import. GracenotesT § 00:18, 13 January 2008 (UTC)

Decoupling the full-date autoformatting and linking systems

This long-standing and vexed issue was raised as a bug at Bugzilla early 06.

After little progress, I presented a petition of 88 Wikipedians here and at Bugzilla in December 06. The petition framed the issue in terms of the greater difficulty of reading, the degraded aesthetic appearance, and the dilution of high-value links caused by the bug.

After further debate, the issue became semi-dormant, but has recently gained momentum at Bugzilla; perhaps this has been in response to moves at MOSNUM to make the autoformatting system optional, on the basis of

  1. the matters raised by that petition,
  2. the perceived inflexibility of autoformatting (in particular, date ranges and slashed dates); and
  3. the realisation that in many articles the vast majority of readers see a mess of inconsistent raw date formats that autoformatting conceals from us logged-in editors.)

The move at MOSNUM to optionalise what many users see as an unsatisfactory functionality has been postponed pending action at Bugzilla. The goal, I think, is to render full dates black, unlinked and autoformatted with a mark-up that is as easy to use/learn as possible. Avoiding back-compatability issues is behind the original proposal to add a new mark-up to the existing system.

Any WPians who have a technical understanding of the issues and who can make useful contributions to the proceedings at Bugzilla are invited to participate. Please keep the discussion focused and as simple as possible, and consider addressing non-expert as well as expert users in your text.

Tony (talk) 07:39, 12 January 2008 (UTC)

Interwiki transfer of edit histories

Problem explained at:

Proposed solution: Just brainstorming, but I suppose a possible technical solution would be this: the transfer software adds "en:" (or the appropriate language/project code) before every username listed in the imported edit history. Would that work? Maybe ask Brion? --Francis Schonken (talk) 14:42, 12 January 2008 (UTC)

Shouldn't that problem not be resolved when the single user login is available? --213.155.231.26 (talk) 15:20, 12 January 2008 (UTC)
Yeah, and when my aunt grows a beard she'd be my uncle. (clumsy translation of Dutch proverb), meaning: "when" doesn't help us with a current problem. --Francis Schonken (talk) 15:31, 12 January 2008 (UTC)

Also, when MediaWiki developers don't see a GFDL problem that would need to be solved (they might be right), there are other problems that might need a solution. I was just wondering whether this might give fatally flawed checkuser results? I mean, a Wikipedia containing edits from two unrelated individuals with the same username (only one of which owns the account), might link IP addresses that are totally unlinked, thus creating false positives via checkuser? Or is this technically excluded? --Francis Schonken (talk) 15:41, 12 January 2008 (UTC)

That should be impossible, as Special:Export created pages do not contain checkuser information. Kusma (talk) 18:21, 12 January 2008 (UTC)
OK, thanks, that's a worry less. --Francis Schonken (talk) 19:07, 12 January 2008 (UTC)

Searching for page containing template X and Y

Is there any way to find all pages that contain two templates? In my case, I want to find image pages that contain both {{Non-free logo}} and {{di-disputed fair use rationale}} (might have parameters) --Apoc2400 (talk) 17:08, 12 January 2008 (UTC)

If both templates have associated categories, you could try CatScan. It might even work if only one template has a category. Anomie 18:00, 12 January 2008 (UTC)
Wikipedia:AutoWikiBrowser has a list comparer function which will allow you to load the transclusions of both templates and find which pages contain both (I don't think you need to be approved to use this function). Let me know if you want some help or if you want me to post the list somewhere. mattbr 22:16, 12 January 2008 (UTC)

Searching Wikipedia

I hope this is the right place for this. When searching the German-language for an article which doesn't exist, one gets this result (screenshot)

German search result

. I think the inclusion of links to immediately search the wikipedia using a variety of external search engines is vastly superior to the results one gets here. Can this be implemented? DuncanHill (talk) 22:25, 12 January 2008 (UTC)

I would certainly support the implementation of this.--Phoenix-wiki 22:35, 12 January 2008 (UTC)
We already have a drop-down menu of search options plus separate links to searchs by google, yahoo and windows live. I suppose the German version is laid out better. 131.111.8.96 (talk) 23:31, 12 January 2008 (UTC)
The German layout is much clearer and easier for the user. DuncanHill (talk) 23:33, 12 January 2008 (UTC)

See MediaWiki talk:Googlesearch#Images: we're less lax than the German Wikipedia on logo use. You also have 6 textboxes for one form submission, not really the best design. On the other hand, our drop-down box requires JavaScript, which limits accessibility a bit. GracenotesT § 00:34, 13 January 2008 (UTC)

de.wp form is also created with JavaScript, and I agree that 6 input text boxes do not make much sense ∴ AlexSm 02:09, 13 January 2008 (UTC)
The inputs are filled in for you. DuncanHill (talk) 02:28, 13 January 2008 (UTC)
In terms of overall user interface design, it's duplicity of function. This makes sending an altered query to separate search engines difficult. We have links to immediately search other engines, by the way. GracenotesT § 17:15, 13 January 2008 (UTC)

Automatic edit summary for interwikis

Most of the anonymous adding interwikis never add edit summaries. While it is not the same as a redirect, blanking or creating a new page, would it be possible to have an automatic edit summary every time one is added? Thanks. -- ReyBrujo (talk) 16:15, 13 January 2008 (UTC)

Most of the anonymous editors never add edit summaries, period. But interwikis are especially annoying in that respect, because anons can't mark edits as minor, and they usually leave out a summary. I agree that it would be very good, if adding or removing an interwiki were all that happened, it would be good to have an automated edit summary. This will probably have to be entered into Bugzilla as an enhancement. Tuvok[T@lk/Improve] 20:15, 13 January 2008 (UTC)
T14613 filled then. I doubt it is trivial (sometimes users don't just insert but also sort or remove other interwikis), but we can hope. -- ReyBrujo (talk) 20:28, 13 January 2008 (UTC)

Hello

I just wanted to say that I believe Template:Keypress needs an actual image of a key. I created a self-made one here. I want to know how to put words inside the actual picture...–Sidious1701(talkemailtodo) 21:37, 13 January 2008 (UTC)

Works fine for me. {{Keypress|Press me. Please.}} produces: Press me. Please. -- Boracay Bill (talk) 00:32, 14 January 2008 (UTC)
I think it should stay the way it is. An image wouldn't scale as well as the current method. For example, the key width required for A versus that for Press this button is huge, and not easily scaled with an image. More realistically, even ⇧ Shift or ↵ Enter are five times the character width of A. RESOLVED WORKSFORME/WONTFIX, I think. Tuvok[T@lk/Improve] 00:55, 14 January 2008 (UTC)

Creating blank page

How does one create a blank page like seen here when there is no deletion log for that page? Quin 11:34, 13 January 2008 (UTC)

Perhaps the first revision was oversighted, and the IP user was affected by this revision or knew that it violated Wikipedia's policies. That's the only explanation I can think of. Graham87 13:22, 13 January 2008 (UTC)
There are several ways it can be done. Subst'ing in a blank template is the most obvious one. If you really, really want to create a blank page, MediaWiki won't stop you; it's just not something that's usually done, and so the software prevents you from creating a page with a completely blank content (assuming it was a simple mistake). AmiDaniel (talk) 09:02, 14 January 2008 (UTC)

User/Talk page edits on blocked IP

I think that it should be possible to edit your own User/Talk page (IF you are logged in) if you are using a blocked IP address. Why can't we? (This is under Technicle because it might be a server bug) —Preceding unsigned comment added by Chirp Cricket (talkcontribs) 07:09, 14 January 2008 (UTC)

Special:Export Parameters... Not working?

I'm part of the Warhammer 40,000 Project, and we're trying to transwiki many of our articles to a Wikia dedicated to Warhammer. We have access to Special:Import on the target wiki, but I'm having problems with Special:Export here on Wikipedia. What I'd like to do is dump a whole article to a file for import, but I can not even begin to get started.

Currently Special:Export gives you the option of the most recent version, or the 100 first edits (so articles that were edited heavily will NOT have an edits newer than the first 100 when exported). I read Manual:Parameters to Special:Export, but neither the links provided at the bottom to export parts of the Main Page, nor trying with my own parameters works. Whatever I do, I get the latest edit (and only that edit) for any article I attempt to use it on and nothing else.

To sum up:

Working

Not working

  • action using it or not seems to make no difference
  • dir can't seem to get it to do anything
  • offset doesn't allow me to select which revisions to export
  • limit I only ever get one version

According to the Manual, this should give me the first five edits of my user page: http://wiki.riteme.site/w/index.php?title=Special:Export&pages=User:Falcorian&offset=1&limit=5&action=submit It does not however, both the link I have in working and Not Working return the same thing, which is the most recent version of my User page.

If someone can explain how to get it to work as documented it would be great because I can then write a script to grab and cat all versions of a page. Also, if anyone has a suggestion on how best to copy the articles if Special:Export won't work, that'd be great too. Thanks! --Falcorian (talk) 07:44, 14 January 2008 (UTC)

curonly=0. AmiDaniel (talk) 08:58, 14 January 2008 (UTC)
Doesn't seem to work, for example: http://wiki.riteme.site/w/index.php?title=Special:Export&pages=User:Falcorian&limit=5&curonly=0&action=submit still produces only the current page. As does http://wiki.riteme.site/w/index.php?title=Special:Export&pages=User:Falcorian&offset=1&limit=5&curonly=0&action=submit --Falcorian (talk) 17:13, 14 January 2008 (UTC)
I thought maybe, just maybe, it was Firefox messing up the php... So I tried wget. Same results, only the most recent version is grabbed, regardless of flags (curonly or otherwise). --Falcorian (talk) 17:52, 14 January 2008 (UTC)
Same problem. Prodego talk 02:15, 15 January 2008 (UTC)

Problem with a wikitable

I'm trying to introduce soft sections into this table but somehow the fourth column, for original air, appears in the Description column and the pipe is visible. I've had a look at the wikitable help but can't see what is wrong here. Can anyone assist? Thanks--Rodhullandemu (Talk) 13:12, 14 January 2008 (UTC)

Help:Table#Pipe syntax tutorial says || is for use on the same line. I think this only works if it's still the same line as the original |. Just replace || with a single | on a new line.[2] PrimeHunter (talk) 15:12, 14 January 2008 (UTC)

I wrote an article called Dissocial personality disorder. However, if I type "Dissocial" or "Dissocial personality disorder" it does not show up in Wikipedia search. I'm wondering if there is a redirect I can't find or some reason why this is happening. Thanks, Mattisse 20:35, 14 January 2008 (UTC)

It certainly works for me. - John Russ Finley (talk) 21:08, 14 January 2008 (UTC)

The search index may take a few days to update with the new article content. Seeing as it's still very new, it may not have updated yet. • Anakin (contribscomplaints) 21:17, 14 January 2008 (UTC)

Yes, and redirects don't show up in search results. It must still have been a redirect when the search index was last updated, so searching on the title doesn't work either. PrimeHunter (talk) 21:27, 14 January 2008 (UTC)
So wait a few days? The article was started from a redirect on January 12. So maybe enough time has not elapsed yet. Thanks! Mattisse 21:33, 14 January 2008 (UTC)

Apologies if this has already been asked, but when I perform an External Links search as here, only the first 50 results are being returned, with no links or options to expand the list to 100, 200, 500, etc. Is there a bug with this tool? - John Russ Finley (talk) 19:52, 14 January 2008 (UTC)

That's very odd. It was working for me last week. But I notice that a new version of the Linksearch extension was released just a few days ago, and maybe was just installed at Wikipedia. It does sound like a bug; not sure who to ask about it though..... • Anakin (contribscomplaints) 20:15, 14 January 2008 (UTC)
Note that it's only the page/limit controls that aren't appearing, the numbers can still be changed in the url, e.g., [3]. • Anakin (contribscomplaints) 20:19, 14 January 2008 (UTC)
Seems to have been fixed. Algebraist 00:24, 15 January 2008 (UTC)
Was reported here. Brion Vibber came to the rescue. :) • Anakin (contribscomplaints) 14:00, 15 January 2008 (UTC)

wikEd as a gadget

wikEd is a full-featured Wikipedia-integrated text editor that adds enhanced text processing functions to edit pages. Currently it works only for Firefox and other Mozilla browsers. wikEd is already a gadget on the German, French, Hungarian, and Occitan Wikipedia as well as the Polish Wiktionary. I suggest to add it as a gadget on the English Wikipedia too and would like to hear your opinion. Сасусlе 05:00, 15 January 2008 (UTC)

I'd support it. wikEd's great, especially for really long articles. Tuvok[T@lk/Improve] 07:12, 15 January 2008 (UTC)
I don't use it much since Firefox isn't my main browser, but it's pretty freakin' cool. I'd support it as well. -- Ned Scott 07:14, 15 January 2008 (UTC)
I definitely support this. It's an excellent script, and it's really no big deal that it doesn't work in non-Mozilla browsers; Firefox is easy enough to install and works better in general with Wikipedia anyway. This is exactly the sort of script that should be a gadget. Pyrospirit (talk · contribs) 15:12, 15 January 2008 (UTC)
I tried it out for some time, and while I am an old-fashioned dude (who writes programs in vim and uses the command line to compile them), I found it good enough, even if a tad slow. -- ReyBrujo (talk) 18:02, 15 January 2008 (UTC)
Now there is. :) Prodego talk 02:04, 17 January 2008 (UTC)
Yes, I agree.--Phoenix-wiki 20:38, 15 January 2008 (UTC)

I disagree, because it does not work in non-mozilla browsers, is quite slow, and is generally bloated. It is extremely difficult to create a WYSIWYG editor entirely in javascript, and I admire the great job that has been done, but I don't think any such implementation will be stable enough and universal enough to include as a selectable option. —Preceding unsigned comment added by Prodego (talkcontribs) 22:45, 15 January 2008 (UTC)

wikEd does not implement any basic text editing function in javascript, it merely sets up the browser-internal rich-text editor. wikEd is also intentionally not a WYSIWYG editor, see here. There is currently no requirement for cross-browser compatibility as long as compatibility is clearly stated, see here. Сасусlе 06:02, 16 January 2008 (UTC)
I have just created the page Wikipedia:Gadget proposals to discuss the addition of new Wikipedia:gadgets. Сасусlе 03:48, 17 January 2008 (UTC)

Main Page is broken?

There seems to be something wrong with the main page. When I click 'view source' instead of opening up, my browser tries to download it. This happens with FireFox as well as IE and also with a friend with a completely different computer and ISP. If I open up the downloaded file I see the source and stuff, as expected. This doesn't happen with 2 other protected pages I tested nor the main page cascade protected T:ITN or Main Page/2 Nil Einne (talk) 08:47, 16 January 2008 (UTC)

Content-Type:	application/x-external-editor; charset=utf-8
I get the same here. but /wiki/Main_Page?action=edit works fine. --Splarka (rant) 09:00, 16 January 2008 (UTC)
(ec) This is strange. I can only edit the main page via http://wiki.riteme.site/w/index.php?title=Main_Page&action=edit&internaledit=true -- does that also work for viewing the source as non-sysop? Kusma (talk) 09:01, 16 January 2008 (UTC)
Yes Nil Einne (talk) 09:02, 16 January 2008 (UTC)
Content-Type:	application/x-external-editor; charset=utf-8
X-Cache:	HIT from sq38.wikimedia.org
X-Cache-Lookup:	HIT from sq38.wikimedia.org:3128
Age:		2146
X-Cache:	HIT from knsq1.knams.wikimedia.org
X-Cache-Lookup:	HIT from knsq1.knams.wikimedia.org:3128
X-Cache:	MISS from knsq4.knams.wikimedia.org
X-Cache-Lookup:	MISS from knsq4.knams.wikimedia.org:80

Apparently someone with the external editor recently viewed the main page source, and caused that specific edit link to become cached (any other edit link worked fine). This shouldn't happen... except apparently the ExternalEdit condition doesn't send any cache control headers, so the squids felt free to cache it. This probably became possible due to some cache refactoring done in the last few days -- from #wikimedia-tech. --Splarka (rant) 10:03, 16 January 2008 (UTC)

I've added a cache-control header. --Brion VIBBER (talk) 21:00, 16 January 2008 (UTC)

User script for automatic translation of non-English webpages in article external links

I've written a user script that will add links to automatic translations of links marked with Language icons. It's at User:AnonEMouse/translators.js. To use it, add the lines:

// Automatic translation of external links
importScript('User:AnonEMouse/translators.js');

to your User:<username>/monobook.js page, then hit shift-reload. It's my first user script, so I may well have screwed something up, use at your own risk, etcetera, but please do report any bugs.

Err... Say what?

For example, without this script, the top of the Argentina#Bibliography section looks like:

To me, with this script, it looks something like:

Clicking on any of the small links will open the named automatic translator's translation of the respective big link into English (in another tab or window for the ones created by my script). Languages supported by at least some automatic translator include: (Arabic), (Catalan), (Chinese), (Traditional Chinese), (Danish), (Dutch), (French), (German), (Greek), (Italian), (Japanese), (Korean), (Portuguese), (Russian), (Spanish), and (Swedish). Translation quality varies, and is rarely great, but for the more popular languages, like Spanish or Japanese, you will at least get several translations to compare, and can usually get some sense of what the translated page is talking about. --AnonEMouse (squeak) 20:00, 16 January 2008 (UTC)

ArticleViewHeader

Note: copied from Wikipedia talk:About the Sandbox -- John Broughton (♫♫) 21:09, 16 January 2008 (UTC)

I heard that on the HRWiki, their sandbox uses the ArticleViewHeader hook to make the sandbox have no header when editing the page. It uses a page called MediaWiki:Sandbox as the header, and if I'm not mistaken, then if we create that same page here, it will make the header show up without it actualy being transcluded on the page. I think it is worth a shot. Soxred93 has a boring sig 19:02, 16 December 2007 (UTC)

Doesn't seem to be any risk in an admin going ahead and being bold. It would be nice to not have to depend on a template that often is erased to give instructions to those editing the sandbox. -- John Broughton (♫♫) 00:19, 9 January 2008 (UTC)
Does that require extensions to WikiMedia not implemented in Wikipedia? If so, an admin won't be enough, you'll need a Developer. --Thinboy00 @180, i.e. 03:19, 10 January 2008 (UTC)
Well, the page mw:Sandbox is in fact a real (editable) sandbox, and the header for it comes from mw:Template:Please leave this line alone and write below (this is the coloured heading), so the above suggestion isn't going to be trivial to implement. But it is true that HRwiki, which uses MediaWiki software, does have a protected (and invisible) sandbox header, so either they have done local customization or there is an extension out there that could be implemented (or something else). -- John Broughton (♫♫) 21:09, 16 January 2008 (UTC)
I don't see anything special here.↔NMajdantalk 21:29, 16 January 2008 (UTC)
I do - there are a bunch of things that were done by a local user there, so that tells me something, http://www.hrwiki.org/index.php/User:It%27s_dot_com confirms it is a local change. You may wish to talk to him. —Random832 00:00, 17 January 2008 (UTC)
It's a really simple extension. The source code is at User:It's dot com/ProjectStaticTopText. Feel free to use it. If it were installed, you would then create MediaWiki pages (or alternatively hard-code default messages) that correspond to the project pages where you want the verbiage to appear. (I don't have any idea whether it its current form it would be appropriate for a wiki of Wikipedia's size. I wrote it a long time ago and it's never given us any problems.) Cheers. — It's dot com 02:19, 17 January 2008 (UTC)

Requested table on articles within a category

Hi. Who would be able to generate for us a statistical table that shows key aspects about the articles in a category? The table would show, for any applicable articles within the category: article name, type of page protection (if any), number of reverts (within last 3 days, if any), 3RR requests (if any), and RfC or 3PO (if any). Please let me know on my Talk if you might be able to create the table. (FYI, the categories sought include Category:Israel and Category:Palestine, for this Wikiproject.) Thanks. HG | Talk 23:32, 16 January 2008 (UTC)

Here's an example of what we'd be looking for, with dynamically generated wiki statistics: battleground statistics for a category. Thanks. HG | Talk 03:53, 17 January 2008 (UTC)
Sounds like something for a bot - which would mean posting at WP:BOTREQ. -- John Broughton (♫♫) 04:30, 17 January 2008 (UTC)

Suggestion for new gadget

I strongly recommend that the script at User:Outriggr/metadata.js be made into a gadget. The script displays an article's assessment rating just below the page title and colors the page title depending on the quality of the article.

This is an ideal candidate to become a gadget available from Special:Preferences. The criteria for gadgets are:

  1. Gadgets must work if just included with no further configuration. (They may be configurable via monobook.js, but must work even if not configured.)
    Requires no configuration.
  2. Gadgets must work in all major browsers (or be clearly marked as to requirement)
    Works in IE7, Firefox 2, Safari, Konqueror 3.5.8, and Ubuntu. Probably works in other major browsers, too.
  3. Gadgets must not be so powerful that new users using them would be likely to cause disruption by mistake.
    Affects display of pages only, so this doesn't apply.
  4. Gadgets must be in MediaWiki space. (This is a technical restriction, although it exists for good reason; it can be fixed simply by moving or copying scripts to MediaWiki space, although the script's maintainer, if any, should be notified.)
    Will need to be copied to MediaWiki space.
  5. There must not be multiple gadgets with exactly the same apparent effect.
    There isn't any similar gadget right now.
  6. Gadgets should not be a collection of multiple scripts (although they can be one script with multiple functions); splitting off the scripts separately makes more sense.
    Has only one function.
  7. Gadgets should not rely on non-Wikimedia websites for their operation.
    Does not rely on other websites.
  8. Gadgets should be useful for a wide set of users.
    Useful for everyone who wants to use Wikipedia for research or get a basic idea of the quality of an article. I've mentioned the article assessment system to many people who were previously skeptical of using Wikipedia, and they all found it very interesting and useful. However, this script currently requires manual installation, which is too complicated to expect it to be widely used. This script really adds a whole new level of reliability to Wikipedia for those who don't know about the assessment system.
  9. Gadgets requiring any permissions other than a username should be clearly marked.
    Requires no permissions.

If this script was available through Preferences, a lot more people would be able to use it, and I really see no reason not to include it.

I'd like to get a general idea on whether this is a good idea or not. Are there any problems with it that I've missed or reasons why it shouldn't be a gadget? It'd also be useful if someone could check to see if this works in other major browsers. Pyrospirit (talk · contribs) 17:19, 13 January 2008 (UTC)

Yes please! DuncanHill (talk) 17:27, 13 January 2008 (UTC)
Just want to add - seems to work fine in Safari on WinXP. DuncanHill (talk) 01:21, 14 January 2008 (UTC)
Works in Konqueror (version 3.5.8 on Kubuntu 7.10). I also think it's useful to be able to see the article class at a glance. NF24(radio me!) 01:16, 14 January 2008 (UTC)
I would rather have some more input on this one, as far as if anyone actually uses that rating system. If so, then I will add it, but I think it would be best to get some more support first. Prodego talk 01:22, 14 January 2008 (UTC)
Works under Ubuntu + Firefox 2.0.0.11, and is a pretty good gadget, because you immediately notice the status of the article. I would like having the option to turn colors off, but I don't object its usage (in fact, I will keep it turned on for the time being). -- ReyBrujo (talk) 01:32, 14 January 2008 (UTC)
By the way, it should also work for diffs, since I rarely visit articles through name, but instead through diffs. -- ReyBrujo (talk) 01:35, 14 January 2008 (UTC)
And a Good article on hold should be treated like Good article candidate. -- ReyBrujo (talk) 02:32, 14 January 2008 (UTC)
That's rather strange. Don't see why FF1.5 should have any trouble with it.

If you want it to work for diffs and oldids, we could use my version at User:Pyrospirit/metadata.js. It's identical to the other script except that I changed a conditional for the begin function from !location.pathname.match('/w/index.php') to wgAction=='view'. Or someone could just edit the gadget version to make this change once we copy it over to the MediaWiki namespace. Pyrospirit (talk · contribs) 03:12, 14 January 2008 (UTC)

So, if no one objects to this, could an administrator please copy the script to the MediaWiki namespace and make it into a gadget? Pyrospirit (talk · contribs) 00:02, 15 January 2008 (UTC)

I haven't tried your modified version, which is the one I would prefer. You should have post it earlier, most people likely tried the other. -- ReyBrujo (talk) 01:54, 15 January 2008 (UTC)
My modified version is identical except that it works any time you're viewing the page content, not just when the URL is in the /wiki/ format. This makes it include diffs and oldids. Since this is such a minor change to the script's functionality, it is irrelevant to whether it should be a gadget or not; an administrator can simply make the change after moving it to the MediaWiki namespace. So… we can probably just proceed to make it a gadget and then make this minor change afterwards. Pyrospirit (talk · contribs) 04:04, 15 January 2008 (UTC)
I have just created the page Wikipedia:Gadget proposals to discuss the addition of new Wikipedia:gadgets. 03:48, 17 January 2008 (UTC)

Discussion continued at Wikipedia:Gadget proposals#Metadata script as a gadget. Pyrospirit (talk · contribs) 14:27, 17 January 2008 (UTC)