Jump to content

Wikipedia:Village pump (technical)/Archive 53

Add links
From Wikipedia, the free encyclopedia
Village pump (technical) archive  

This page contains discussions that have been archived from Village pump (technical). Please do not edit the contents of this page. If you wish to revive any of these discussions, either start a new thread or use the talk page associated with that topic.

< Older discussions · Archives: A, B, C, D, E, F, G, H, I, J, K, L, M, N, O, P, Q, R, S, T, U, V, W, X, Y, Z, AA, AB, AC, AD, AE, AF, AG, AH, AI, AJ, AK, AL, AM, AN, AO, AP, AQ, AR, AS, AT, AU, AV, AW, AX · 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43, 44, 45, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 57, 58, 59, 60, 61, 62, 63, 64, 65, 66, 67, 68, 69, 70, 71, 72, 73, 74, 75, 76, 77, 78, 79, 80, 81, 82, 83, 84, 85, 86, 87, 88, 89, 90, 91, 92, 93, 94, 95, 96, 97, 98, 99, 100, 101, 102, 103, 104, 105, 106, 107, 108, 109, 110, 111, 112, 113, 114, 115, 116, 117, 118, 119, 120, 121, 122, 123, 124, 125, 126, 127, 128, 129, 130, 131, 132, 133, 134, 135, 136, 137, 138, 139, 140, 141, 142, 143, 144, 145, 146, 147, 148, 149, 150, 151, 152, 153, 154, 155, 156, 157, 158, 159, 160, 161, 162, 163, 164, 165, 166, 167, 168, 169, 170, 171, 172, 173, 174, 175, 176, 177, 178, 179, 180, 181, 182, 183, 184, 185, 186, 187, 188, 189, 190, 191, 192, 193, 194, 195, 196, 197, 198, 199, 200, 201, 202, 203, 204, 205, 206, 207, 208, 209, 210, 211, 212, 213, 214, 215, 216

Fundraising Banner Bug

Hi all-
I was logged out and I noticed that there seems to be a bug with the fund-raising banner. When I hit the collapse button, it didn't collapse. Obviously, when I logged in it went away (I have it disabled in preferences) but I imagine that would be somewhat annoying to an anonymous user who can't get the giant obnoxious banner to collapse. I'm using the latest version of Firefox under Ubuntu with javascript enabled. l'aquatique || talk 08:34, 12 December 2008 (UTC)

T̂here's definetly something wonky going on with it. On my watchlist, it will collapse only the 'donate now' part at the bottom, and replace that with an 'expand' button, all the while the banner stays, 'collapse' button and all. ♫ Melodia Chaconne ♫ (talk) 12:31, 12 December 2008 (UTC)
Might be a bug in one of the banners in rotation; we'll take a look. --brion (talk) 15:46, 12 December 2008 (UTC)
It was indeed a bug where the main notice div was missing the collapse id it needed. eg. "siteNoticeBig notice-wrapper" not not just "notice-wrapper" All quotes are fixed now and thanks for spotting it so quickly. --Tfinc (talk) 19:36, 12 December 2008 (UTC)

I noticed this on one of the "quotes" banners. The collapsed version appeared without removing the big version. — Werdna • talk 15:54, 12 December 2008 (UTC)

I just noticed this on one of the quotes banners as well. EVula // talk // ☯ // 18:19, 12 December 2008 (UTC)
Yeah, I think it was a quote banner that I was having trouble with. If I remember correctly it was the one with a quote written in the Cyrillic alphabet from someone in Israel? However, I just logged out to check and it seems to be all the quote banners. The one with the old fashioned scale on it collapses properly, though. Also, when you try to collapse it the replacement banner also appears- suggesting to donate to Meta. Aren't our banners supposed to say donate to Wikipedia just as Commons' banners say to donate to Commons? l'aquatique || talk 19:09, 12 December 2008 (UTC)
The 'meta' thing is a separate bug, which happens occasionally but we haven't been able to track it down yet. (The notices are actually generated through the Meta database, but they override the language and sitename while doing the generation. So it *shouldn't* be possible for it to say 'Meta'... but sometimes something goes awry. :( ) --brion (talk) 00:43, 13 December 2008 (UTC)

Check the American quote banner, the Donate now and Learn more buttons are a little to the right and down - hanging outside the banner box. --Clark89 (talk) 04:59, 17 December 2008 (UTC)

Can you be a bit more specific about which browser version you are using on which OS and if you are increasing the size of your fonts in any way? I've loaded the English banner with FF 3.04 (OSX) , IE6 & 7, Safari (OSX) and Opera (OSX) and they all display fine with their defaults. The only potential problems I can see are if your screen size is 800x600 or smaller then the learn more gets pushed down and/or you have increased the size of your fonts from the defaults. Do let know --Tfinc (talk) 18:51, 17 December 2008 (UTC)
No problems with FF (Ubuntu) at screen res 1280x1025 normal font size. l'aquatique || talk 19:10, 17 December 2008 (UTC)
OS- WindowsXP. Internet browser- IE7. Screen- 1024x768. Changing the text size in the browser from medium to small yields all contents inside the banner, but still slightly off position of where the other banner's buttons are. Couldn't find any relevent settings enabled that might cause problem. --Clark89 (talk) 22:50, 17 December 2008 (UTC)
The second American banner and the Italian banner are doing it too, the American banners have the "Learn more" straight down-over the line about 10%-the "Donate now" 50% over the line to the right, the Italian banner "Learn more" 90% over the line. At 1280x1024 the "Learn more" buttons are a bit above where the others would be and the "Donate now" well inside the box. --Clark89 (talk) 06:02, 22 December 2008 (UTC)

null edits

Is there a bot I can task with doing some null edits for me? --Pascal666 (talk) 03:44, 19 December 2008 (UTC)

Requests can be posted to Wikipedia:Bot requests but which null edits do you have in mind? Things done by a null edit should also be done automatically at some time. They are just waiting in the job queue and there is often no compelling reason to stress the servers by mass-expediting them with null edits. PrimeHunter (talk) 03:57, 19 December 2008 (UTC)
I added a category to a template about 35 hours ago. None of the page categories have been updated yet. I just tested a null edit and it updated the page just fine. This would tend to indicate the template is working fine, but the job queue is not. --Pascal666 (talk) 04:11, 19 December 2008 (UTC)
Some things just take time. Are you referring to Category:Pages that use the Sockpuppet category template incorrectly in [1]? PrimeHunter (talk) 13:18, 19 December 2008 (UTC)
Yes, sorry, didn't expect you to go hunting for it. Does appear to be working now. Guess I'm just not used to the job queue being two days behind. Thank you for your time. --Pascal666 (talk) 21:35, 19 December 2008 (UTC)

The point of using a bot account is to avoid flooding recentchanges with high-speed edits and to have a higher line limit when scraping from the API. Since these don't apply (nobody sees the null edits anyway) there's no benefit to using a bot account. — CharlotteWebb 01:29, 22 December 2008 (UTC)

The question was not about using a bot account. The question was about having a bot (a piece of software that can automatically make [null] edits to a list of pages) do the edits. The benefit to having a bot do it is not having to manually edit several thousand pages. --Pascal666 (talk) 04:53, 22 December 2008 (UTC)

Script issue

Any recent changes (say, the past 12 hours or so) that have affected .js scripts? This morning, I've noticed that the tabs for "Friendly" and several other scripts that I load from monobook.js are not appearing as per usual (especially on user/user talk pages). I've already reloaded Firefox, and I haven't added anything new to monobook.js recently. Thoughts? --Ckatzchatspy 17:22, 19 December 2008 (UTC)

Check whether it is flooding your error console, might be some useful information there. — CharlotteWebb 01:23, 22 December 2008 (UTC)
This was happening to me too, it has resolved itself on one computer (Mac OS X 10.4 w/ FF3.0.2) but not on another (Windows XP SP2 w/ FF3.0.2). Foxy Loxy Pounce! 04:37, 22 December 2008 (UTC)

Is there any way to restrict the namespace for a link search? Several times a day, I wind up staring at [this list], and trying to figure out which links have been added. I really don't care if people link to the page from talk space or user space, I only need to search and destroy links from article space. Is there any way to look for the link only in article space?—Kww(talk) 22:03, 19 December 2008 (UTC)

Special:LinkSearch has support for limiting the search by namespace, but that feature is disabled here. You could use the API to do the search (like this), but the output is less user-friendly. Anomie⚔ 23:38, 19 December 2008 (UTC)
A lot more user-friendly than going blind looking at that list. Thanks. Why on earth is this functionality disabled?—Kww(talk) 00:57, 20 December 2008 (UTC)
You'd have to ask the sysadmins. The comment on the configuration entry in CommonSettings.php is "wgMiserMode now in the configuration object enabled to test DB load balancing -- TS 2004-06-22". Anomie⚔ 02:09, 20 December 2008 (UTC)
It's a very inefficient query with the namespace limiting in. --brion (talk) 22:48, 21 December 2008 (UTC)
Just curious, if namespace selection was enabled, would it be invertible as described in bugzilla:14485? — CharlotteWebb 01:15, 22 December 2008 (UTC)

Searching edit summaries

Is there any way to search my edit summaries? I ask because I do a lot of dabbing of Georgia, and I'd be interested to know just how many I've done. DuncanHill (talk) 05:40, 20 December 2008 (UTC)

You can use the API to get a list of your edit summaries in 500-run batches, copy them into Excel or Word, and use Find to pick out key phrases. That's usually what I do if I want rough numbers on things like this. Happy‑melon 21:50, 21 December 2008 (UTC)
You could also look at your contributions and use your browser's Find feature to search for a specific phrase.
Somewhat related, if you want to see your stats for a specific page, there's a tool for that.[2] EVula // talk // ☯ // 05:22, 22 December 2008 (UTC)

Logs

We need to make the log page have more search options. For example here, we need to be able to search for users that don't have any talkpage comments, users that never had any contributions, and we want to search for blocks with expiry times of less that 3 days, or more than 1 week but less than half a year, etc., and we want to search for, or exclude (like the boolean operator "NOT", certain "block summaries" (such as "Spamming promotional material", "Vandalism", "Intimidating behaviour/harassment", "Inserting false information", etc. etc.).96.53.149.117 (talk) 00:51, 22 December 2008 (UTC)

Why? What purpose would it serve? — The Hand That Feeds You:Bite 01:13, 22 December 2008 (UTC)
Searching block/delete reasons wouldn't be reliable because not everyone uses the same words to mean the same thing, or spells them correctly. — CharlotteWebb 01:19, 22 December 2008 (UTC)
You can download the logs from October 2008 in SQL format (522.6 MB). Graham87 06:37, 22 December 2008 (UTC)
Note that that's not the logs from October, that's the logs up to October. Mr.Z-man 20:13, 22 December 2008 (UTC)

Another Static site question

To go along with my last (answered, thanks) question; is there a static site for the partner programs, like wikinews and wikisource that I could download compressed static images from like the static wikipedia site? Not that I'd be reading it any time soon, but I'm still interested. Lostinlodos (talk) 09:34, 22 December 2008 (UTC)

Not at this time... --brion (talk) 17:22, 22 December 2008 (UTC)

Category:Chief executives : Vandalism

Can someone :

  • restore (after vandalism) the former version :

23:54, 28 October 2008 Kal-El-Bot (Talk | contribs). I do not know how to do.

Thanks

Eras-mus (talk) 22:18, 22 December 2008 (UTC)

Both done. Prodego talk 22:23, 22 December 2008 (UTC)

Inline citations for web pages

I need some help with inline citations. Please see my posting at Talk:Peter Pan (1954 musical)#Adding inline citations. Thanks. Thomprod (talk) 18:53, 19 December 2008 (UTC)

Dendodge and I have responded at Talk:Peter Pan (1954 musical)#Adding inline citations. – ukexpat (talk) 19:15, 19 December 2008 (UTC)
Thanks, will do. Thomprod (talk) 18:12, 20 December 2008 (UTC)
Resolved
 – Thomprod (talk) 18:54, 23 December 2008 (UTC)

wikiproject template exception processing continues, even after fix

Our Wikiproject template {{WikiProject Oregon}} was affected by the namespace rename of Image: to File:. Changes on December 15,2008 should have fully addressed the change, but haven't—at least not completely. Logic in the template puts unrecognized namespace pages in category:Oregon articles needing attention which continues to have lots of entries: 450, about as many as when I first noticed the problem back then. Visiting any of those pages says it is not in the category.

This kinds of seems like the same problem above at #Templates slow to refresh, except ours is much slower to be addressed. Any ideas? —EncMstr (talk) 08:44, 22 December 2008 (UTC)

Null edit to the individual pages fixes them, so it is just a waiting game (damn job queue). --Splarka (rant) 08:08, 23 December 2008 (UTC)

sdfdsf

You will see that the top case, if you put it in the sandbox, will create a line, an extra white line. Now, I won't contribute to the discussion further, but I think it might be a good idea to discuss this and see what the community decides: that we should take this to bugzilla so that a line below __TOC__ will not wysiwug, so that these lines will only function as editingease.96.53.149.117 (talk) 23:17, 22 December 2008 (UTC)

I changed the above comment to use <pre> tags rather than <nowiki> tags, so that the formatting mentioned is readily readable outside of the edit mode. {{Nihiltres|talk|log}} 23:58, 22 December 2008 (UTC)
Apologies if I'm missing the point of this concern. If someone cares about making sure whitespace is absolutely perfect on a page, the thing to do is look at the results and tweak it until it comes out right. A change via Bugzilla might undo the work of everyone who has already done that. --A Knight Who Says Ni (talk) 03:09, 23 December 2008 (UTC)
Doesn't need to be fixed via MediaWiki, IMO. It works as intended; in some cases, people want those extra lines. Gary King (talk) 20:06, 23 December 2008 (UTC)

Image does not display, rather attempts to download

When I click on the image File:"City of India,Actress Prachee Adhikari for some promotion,New Delhi 1 - Oct 2007. jpg to display at full screen resolution, it does not display it in the browser as with other images, but rather attempts to download it on my machine. Is this a bug in Wikimedia software, or is it on my browser (Firefox 3.0.5, Ubuntu x86_64)? And even if it is my browser, it seems likely this is a common problem to Firefox; so should it be fixed? Magog the Ogre (talk) 07:43, 23 December 2008 (UTC)

Well, that's a bug somewhere. Either in the software for allowing uploads with spaces in the extension, or in the servers for serving them as text/plain. --Splarka (rant) 08:15, 23 December 2008 (UTC)
Do you mean quotations? There are no spaces; they are underlines (_) as they normally are. Magog the Ogre (talk) 08:33, 23 December 2008 (UTC)
Splarka is right; but to clarify, the problem is the image has an extension of . jpg (note the space). This shouldn't have got that way, and the simplest thing to do is to save the file and remove the space from the extension using your OS. This, that and the other [talk] 09:27, 23 December 2008 (UTC)
Can somebody fix this image? I know it's under license review, but as I said the filename shouldn't be that way. It needs to be renamed. This, that and the other [talk] 09:37, 23 December 2008 (UTC)
It's under review, unused, and an obvious copyright violation. Unless you have the ability to erase the old one, it might be best to do just do that straight out; I don't see any point in that, especially if any programmers can look at the bug shortly. Magog the Ogre (talk) 12:57, 23 December 2008 (UTC)
I've filed it as bugzilla:16772; we should have it patched up pretty soon. Feel free to delete the file at your leisure. --brion (talk) 17:17, 23 December 2008 (UTC)
 Done Painfully obvious copyvio. EVula // talk // ☯ // 01:41, 24 December 2008 (UTC)

Mediawiki version that Wikipedia is currently using

I'm trying to find (again) the version of Mediawiki that Wikipedia is currently using. I found a categories bug that I've already reported and I was wanting to know how current Wikipedia is. About once a year or so, I find some bug and need to know the version and every time I have trouble finding this information. This time I found several pages that used to have it but are obsolete. So, what is the best place? And are there several pages that list it? Jason Quinn (talk) 15:48, 23 December 2008 (UTC)

Have you tried Special:Version? Mayalld (talk) 15:50, 23 December 2008 (UTC)
Thank you. Nope... I found about 10 different versions of using "Special" and "Version" but not the obvious one. Thanks for the quick reply. Jason Quinn (talk) 15:53, 23 December 2008 (UTC)
1.44.0-wmf.2 (8fd6c9c) (<- version automatically updates). Dendodge TalkContribs 15:59, 23 December 2008 (UTC)
Cool. Learned something new. Thanks guys. Jason Quinn (talk) 16:04, 23 December 2008 (UTC)

MediaWiki updated

I've updated MediaWiki on the cluster to r44990. List of changes... :) --brion (talk) 23:53, 23 December 2008 (UTC)

Display problem on diff pages

See the Difference between revisions part. Has the recent update done this? Or has someone messed with the MediaWiki page? Foxy Loxy Pounce! 00:28, 24 December 2008 (UTC)

No, the MediaWiki page (MediaWiki:Difference) is fine (or so I think). The update may have done something... Calvin 1998 (t·c) 00:34, 24 December 2008 (UTC)
According to the release notes linked by Brion a number of MediaWiki messages now use wiki syntax instead of HTML. The specific message in your image cap is MediaWiki:Difference (edit | talk | history | links | watch | logs). —Locke Cole • t • c 00:36, 24 December 2008 (UTC)
Update requested. Calvin 1998 (t·c) 00:40, 24 December 2008 (UTC)

MediaWiki message changes

The following messages were affected by the update and may need to be fixed/edited.

—Locke Cole • t • c 00:42, 24 December 2008 (UTC)

Proposal to carve off editprotect as a separate access level

There is a policy proposal, Wikipedia:Village pump (policy)#New individual access level: editprotected, that is getting significant support. Before I take it to Wikipedia:Requests for comment/Policies I want some technical feedback: Is this relatively easy to do if the community requests it? davidwr/(talk)/(contribs)/(e-mail) 20:20, 19 December 2008 (UTC)

The spam blacklist isn't protected; it's an interface message, so it can only be edited by people with the editinterface user right. Technically, I believe that right could be unbundled, but you'd need to trust the recipient not to insert malicious code in the wrong place, etc. -Steve Sanbeg (talk) 21:17, 19 December 2008 (UTC)
While this originated out of an SBL-related RFA, various discussions including this one have concluded that spinning off the editprotect permission would have a higher chance of gaining consensus while at the same time being useful to the project. That is the permission that will be in any related future RFC. davidwr/(talk)/(contribs)/(e-mail) 01:57, 20 December 2008 (UTC)
OK; well, there is an editprotected right, so it should just be a configuration change to edit protected pages, but that wouldn't affect system messages or pages with cascading protection (neither the source, nor the transcluded pages), to prevent them from using that to protect other pages. -Steve Sanbeg (talk) 14:57, 22 December 2008 (UTC)
It looks like this would be simple to do if there were consensus. —Simetrical (talk • contribs) 18:29, 24 December 2008 (UTC)

Interiot / wannabekate tool should be deprecated?

Based on some recent investigations and discussions, it's looking like perhaps the use of Interiot's edit counter should be deprecated. Based on the stated source code, the recent change of space name from Image: to File: is a definite breakage (per CharlotteWebb) and I believe that there is a flaw in the assignment of namespaces by means of regex'es, evidenced in the counting discrepancies I've found in the help desk thread linked above.

Just putting it out there, if others can confirm, maybe we need to retire Interiot's valuable tool throughout the wiki (unless Interiot feels like updating it, which would be nice too :). Franamax (talk) 02:16, 22 December 2008 (UTC)

AFAIK, they've retired. That's why we have other edit counters.  — Mike.lifeguard | @en.wb 03:42, 22 December 2008 (UTC)

This is not a vote, and not something that where editing a bunch of pages to change the url would do much good. We need to ask the toolserver overlords to disable it and replace it with a list of links to tools which still work, so people can update the bookmarks in their browser, etc. — CharlotteWebb 03:54, 22 December 2008 (UTC)

Voting is evil, what made you think I was voting, I was merely indenting and bolding my point, much like RfA ;) Foxy Loxy Pounce! 04:55, 22 December 2008 (UTC)
I had looked at some other edit counters a while back, but none had the visual aid that Interiot's had; the bars with lengths based on how many edits a user made in a month was very useful. Are there any others with graphs like that? Gary King (talk) 20:08, 23 December 2008 (UTC)
Something like this? Mr.Z-man 20:20, 23 December 2008 (UTC)
Neat, they finally added it. Gary King (talk) 20:23, 23 December 2008 (UTC)
Problem is, those graphs aren't half as easy to read as those on wannabekate's tool, which are great. Is there no way that the existing tool can be updated? TalkIslander 18:54, 24 December 2008 (UTC)
There is also my edit counter, based on Interiot's. Xclamation point 03:37, 25 December 2008 (UTC)

Fundraising banner (Personal appeal)

For me, the 'A personal appeal from Wikipedia Founder Jimmy Wales' banner over-reaches its boundaries, with a line cutting through the characters on the bottom line. This is observed in IE 7.0 (somebody else's computer!). The font also looks too large to be professional... — Werdna • talk 13:02, 23 December 2008 (UTC)

I can reproduce this with fonts cranked up to 120dpi. I'll put Tomasz on it when he gets in to the office (bugzilla:16773) --brion (talk) 17:27, 23 December 2008 (UTC)
It's still overflowing the line, and looks very bad. In the mean time, could it just be disabled in favor of only the foreign quotes and " Wikipedia needs you. Support the largest non-profit project on the web. "? I think the quotes are by far the best "ad" so far. Superm401 - Talk 03:25, 25 December 2008 (UTC)

I seem to have lost the "what links here" link in the side-bar of classic skin (it now appears only at the bottom, rather inconveniently as I'm not in the habit of using those links). "Related changes" has also gone. Is this deliberate? —  Tivedshambo  (t/c) 06:22, 24 December 2008 (UTC)

You refuse to keep up with the times, too? ;-) Folks fiddle around with the parts of the page architecture, test the changes to see that it works okay on the Mono skin ... & the rest of us who prefer other skins are left to shift for ourselves. I've been trying to figure out just what got changed & see if I can't hack the standard.css & standard.js files to restore some of the quickbar features to that skin, but I can't find an "Introduction to Customizing Skins" page just a lot of clues & hints to those who know CSS -- or know how the Wikipedia page layout is put together. -- llywrch (talk) 07:32, 24 December 2008 (UTC)

All pages in a namespace

Resolved
 – Thanks

Is there any way to get a complete list of all pages in the Wikipedia: namespace? --Unpopular Opinion (talk) 18:40, 24 December 2008 (UTC)

Searching for "Wikipedia:*" doesn't help. --Unpopular Opinion (talk) 18:42, 24 December 2008 (UTC)
Looking at File:Namespaces@enwiki.pdf, there are over 300,000 pages in the Wikipedia namespace. I do not know any tool that would easily sort that size of a table. MBisanz talk 18:48, 24 December 2008 (UTC)
Special:PrefixIndex. EVula // talk // ☯ // 18:50, 24 December 2008 (UTC)
More specifically, here's a list. Majorly talk 18:54, 24 December 2008 (UTC)

move=autoconfirmed

When decreasing the protection level of a page, especially with the new separation between edit protection and move protection, "move=autoconfirmed" appears, and it has no effect since moving is restricted to autoconfirmed. Ideally, when saving a page with only this level, it should read it as a complete unprotection. To this effect we may create a bug. In the meantime, it still creates problems with bots, etc. So is there a way to have a list of all those pages, so that we can unprotect them entirely ? Cenarium (Talk) 00:34, 25 December 2008 (UTC)

From an API query and a quick Python script, these are all the pages in ns:0 with only move=autoconfirmed set:
List of pages

--Mr.Z-man 03:58, 25 December 2008 (UTC)

Thank you. They are fixed now. Cenarium (Talk) 17:44, 25 December 2008 (UTC)

See T18791. Cenarium (Talk) 19:18, 25 December 2008 (UTC)

Enwiki database dumps

Dear Sysadmins,

Did you get enough money to make another enwiki dump?

The last one which succeeded, for those of you following along, was October 2006. 69.228.221.245 (talk) 05:39, 25 December 2008 (UTC)

The latest download dump is from October 8th, 2008. You can get it here. —Nn123645 (talk) 06:42, 25 December 2008 (UTC)
The full "all pages w/ complete edit history" dump is still in-progress according to that download page. Calvin 1998 (t·c) 08:06, 25 December 2008 (UTC)
The current dumping code is slow and will fail completely if anything goes wrong. The devs are working on new code that will be faster and will be better able to recover from errors. Anomie⚔ 14:52, 25 December 2008 (UTC)

Templates slow to refresh

I've noticed this before - when I fix a link in a template it can take days for the "What links here" links to refresh. Take these two examples:

On the article Life with Father (film), I split out most of the info from the main Life with Father article, including updating the director template for Michael Curtiz. This was yesterday, however when I check to what links to the film, only the cast & crew I've manually updated do, and all the films in the template still point to the original article. I've something similar with film templates (correcting disambig links, etc), and it's taken an age for the links to seemingly point to the correct article.

The second example is for Lady in the Dark. The talk page has a template asking for work on content forking (Template:Musicals-tasks). I removed the link for LitD, as I'd created the film page a few days ago, but there are still hundreds of talk pages within the Musical-Theatre project that still link to the original article. I know that if I wait a few more days, the changes will reflect the updates on the templates. Any ideas? Lugnuts (talk) 09:37, 21 December 2008 (UTC)

You already know the solution (wait), so I'm not sure what you want here. You can force it with null edits, but that's time-consuming and puts more strain on the servers, and there seems to be no pressing reason this has to be done fast. Algebraist 09:49, 21 December 2008 (UTC)
Yes, I can wait, but I was wondered why it took so long? I wanted to correct any links pointing to the wrong article(s). As it takes so long, its easy to forget which article it affects and not to come back later to check. Lugnuts (talk) 10:17, 21 December 2008 (UTC)
Uh, to reduce the server load? 199.125.109.64 (talk) 20:29, 21 December 2008 (UTC)
You can see the queue length at Special:Statistics. For general background read [3] and [4]. I can't find $wgJobRunRate in any of the config files at [5]. Hopefully Wikipedia has it set to 0 and is using cron (would seem to be supported by [6]). Maybe it is time to add another server to the pool that runs these jobs? --Pascal666 (talk) 20:36, 21 December 2008 (UTC)
Thanks Pascal. Lugnuts (talk) 10:21, 22 December 2008 (UTC)

There's a way to force a refresh to update changes from recently updated templates.

  1. Click "edit this page" on the article where the template is placed.
  2. Notice the page URL having &action=edit at the end. Change that to &action=purge and hit Enter.
  3. Hopefully the template is updated in the article.

Hytar (talk) 21:12, 22 December 2008 (UTC)

Nope, that hasn't worked either! Lugnuts (talk) 09:44, 26 December 2008 (UTC)
Right. Hytar described purging but that generally only affects the purged page. It requires a null edit to update "What links here". PrimeHunter (talk) 21:04, 26 December 2008 (UTC)

Oh yeah... I didn't notice that. Guess I'm a bit general on the subject. Thanks for pointing out! HУтaяtalk2mecontribs 21:48, 26 December 2008 (UTC)

Categories from templates are also updated very slowly recently. While it was done in a few hours some weeks ago, some haven't been updated for more than a week. Cenarium (Talk) 22:56, 26 December 2008 (UTC)

Help us track down Internet Explorer "Page not found" error

Hey all -- one of the new goodies when we updated the site earlier this week is that hits to non-existent pages now return the proper HTTP 404 "Not found" response code. This means that search engines and link checker utilities can now properly detect that a link to a wiki page is bad due to a typo in the link, or the page having been deleted.

When we first tried this in 2005, a number of editors found that this wasn't working right for them in Internet Explorer -- they were getting a generic "Page not found" error page shown instead of the MediaWiki interface. At the time we had to revert it quickly, because the 404 code was applied to edit pages for redlinks as well as straight page views, so affected people couldn't create new pages.

The new version doesn't apply to edit pages, so we figured we could slip it in and debug the Internet Explorer issue at our leisure if it appears again. I have gotten one report (from it.wikipedia.org) but we didn't have enough time on IRC to fully figure out what was going on yet -- I haven't been able to reproduce it consistently on my Windows box and haven't gotten detailed enough data to figure it out.

I need some folks who use Internet Explorer to try a few pages and see if they're experiencing the problem... If you do see the error pages, I'd appreciate it very much if you could try capturing a trace with FiddlerCap or another web debugging tool or packet analyzer and sending it to me at brion@wikimedia.org. (If you're logged in, the trace will include your session cookies so do not post such a trace in a public place!)

Possible steps to reproduce:

If you're seeing Wikipedia's interface, you don't have the problem. If you're seeing a generic Internet Explorer error page, you've hit the bug.

Thanks! --brion (talk) 21:41, 26 December 2008 (UTC)

Can't change protection level for Middle Ages

I must be doing something dumb here. I accidentally fully protected this article (at least I assume that is what happened, as it is acting a bit weird now I'm wondering if it was me after all). Now when I try to either unprotect entirely or change to semi-protect I get the message 'Expiry time is invalid'. Can someone please tell me what I'm doing wrong and change it to semi-protected for me? Thanks. dougweller (talk) 22:00, 26 December 2008 (UTC)

I managed to unprotect it, by fiddling with the settings, though it took me two tries. I don't know what we were doing wrong. ·:· Will Beback ·:· 22:15, 26 December 2008 (UTC)
I changed it back to semi+move without expiry problem. I can't reproduce this. Maybe the initial expiry was partially invalid, mediawiki managed to find out the time but couldn't modify it. Cenarium (Talk) 22:45, 26 December 2008 (UTC)
Thanks, weird. dougweller (talk) 06:25, 27 December 2008 (UTC)
I think you get the "expiry time is invalid" notice when the move or edit level is set to "allow all users", but it has an expiration time. Gimmetrow 06:35, 27 December 2008 (UTC)
I have tried this but it just unprotects in this case. Cenarium (Talk) 15:24, 27 December 2008 (UTC)
I know I've run into "expiry invalid" before, too. Here's a scenario: with the "move" box unlocked, if you have something typed in the "other time" field for "edit" (for either allow all, semi or full), and select "infinite", the "other time" field is not cleared. This will lead to an "expiry invalid" notice. Gimmetrow 17:44, 27 December 2008 (UTC)

Issue with my Subpage- [7]

I need some help from someone of a higher technical grade here, I'm stumped. I created a User Subpage because I was sick of Wikipedia's complex "Welcome to Wikipedia" subpages and how they were scattered across. So I spent a couple of minutes making a page that would give new Users a simple, 5 minute basic introduction that would get them editing in 5 minutes or less, when something strange in spontaneuos happened. I clicked "Preview" for what must have been the 12th time and it at 1/4 of my page. The HTML code was still there for the last quarter, but it simply would not display it. I clicked a few more tiles, and it seemed to continue to eat my work. I also tried copying it out on another page too see if it was a problem with that one URL, but it persisted. I'm clueless here- does anyone have any idea what's happening? If you do, post here or on my talk page If you can help me, please reply! Resident Mario (talk) 01:43, 27 December 2008 (UTC)

I would think seriously about the tone of the page (ignoring for a moment the spelling and grammatical errors) - references to "pesky admins" and "diheart [sic] Wikipedians commited [sic] to minor grammatical errors...", are not doing you any favours. – ukexpat (talk) 02:07, 27 December 2008 (UTC)
(e/c) :You had an unclosed <ref> in your text. --TheDJ (talk • contribs) 02:09, 27 December 2008 (UTC)
You knwo what, forget this, this isn't helping. Sorry I asked. And it's called humor. And so waht about gramatical errors, its a subpage, not a Wiki article.UNGResident Mario (talk) 20:07, 27 December 2008 (UTC)

Monobook

I can't seem to find my monobook.js file. It sends me to some blank yahoo page. Can someone please help. Thank you. --CyclePat2 (talk) 15:25, 24 December 2008 (UTC)

User:CyclePat/monobook.js is what I'm assuming you're talking about. EVula // talk // ☯ // 15:47, 24 December 2008 (UTC)
Or possibly User:CyclePat2/monobook.js. – ukexpat (talk) 15:54, 24 December 2008 (UTC)
Eh, I figured CyclePat was posting under his alt account due to monobook issues. *shrug* EVula // talk // ☯ // 19:02, 24 December 2008 (UTC)
Hi EVula. thank you for the reply. Okay! euh! Sorry. Wasn't working earlier because the page doesn't exist. I think this is something with my IE search engine. Anyways, my search, under Wikipedia's search, would redirect to some yahoo search (that the search engine bar I use). Weird but I guess it makes sense in a way. I guess all I need to do is try it out with firefox and I bet you it would have works. Anyways. Should have though about making a direct link (aformentioned) which we can click on!!! Thank you and Merry Christmas. b.t.w: It was the CyclePat2 account.--CyclePat2 (talk) 00:29, 25 December 2008 (UTC)
I believe the developers have just implemented 404:file not found http codes for non-existent pages. This has the side affect that some browsers will redirect users to "friendly" file not found pages, instead of showing Wikipedia's. I think you can disable "friendly http error messages" in the internet settings and disable third party extensions (toolbars being evil), but Firefox really is the way to go. — Blue-Haired Lawyer 03:47, 29 December 2008 (UTC)

JavaScript

Is it possible to edit Wikipedia through JavaScript? The (i think) standard language hinging <script language= ... type= ...> doesn't work. 24.185.37.213 (talk) 17:42, 28 December 2008 (UTC)

Erm, I don't actually understand what you're asking. JavaScript is a language, not a program or interface. EVula // talk // ☯ // 18:02, 28 December 2008 (UTC)


If you mean 'is it possible to insert arbitrary JavaScript into Wikipedia articles' then no, it is not, because that would be a really terrible idea. Algebraist 18:09, 28 December 2008 (UTC)


1. Exactly. JavaScript is so close to HTML, that HTML tags are interchangable in JavaScript- so i was suprised you couldn't just switch to JavaScript entirely.


2. Wait, why would that be a bad idea? ...oh yeah, I geuss it would get confusing (by the way, I meant user-pages, notarticles, altogh that's a good point). But isn't the Qui program imported JavaScript? Resident Mario (talk) 18:18, 28 December 2008 (UTC)

Users can add custom JavaScript just for themselves in their personal JS subpage and admins can ad JS for the entire site, but allowing anyone to add any JS to a page that would be executed by anyone would be a massive security risk. You can't use most HTML in pages either for similar reasons. Anyway, JavaScript isn't close to HTML at all, I'm not sure what you mean by that. JavaScript is a scripting language, HTML is a markup language. The two do entirely different things. (and I've no idea what "Qui" is, it doesn't sound like anything I've ever heard of on Wikipedia) Mr.Z-man 18:40, 28 December 2008 (UTC)
Furthermore, while JavaScript can output HTML, and while it's technically possible to have the MediaWiki system output JavaScript that in turn outputs HTML, it would be inefficient to such an insane degree that the servers would crash once there were more than five people accessing the site. EVula // talk // ☯ // 19:24, 28 December 2008 (UTC)
Well, it's feasable to deliver a more compact stream of data using JavaScript. A lot of content on the site is already generated at least in part using JavaScript. JavaScript is executed client-side and procedurally, so it wouldn't necessarily cause the servers to "crash". However, it would require a near total rewrite of MediaWiki, and require *yet another* and *even more* confusing API for users to adopt, on top of the existing JavaScript, PHP, WikiMarkup, etc. I think that even AJAX applications eschew using JavaScript to store data (though I've personally used it a few times) in favor of XML. SharkD (talk) 19:59, 28 December 2008 (UTC)
You can also modify web pages dynamically by using JavaScript in the address bar, or by using certain browser extensions (such as Internet Explorer's Developer Toolbar). The changes in these cases tend not to be permanent, however. I haven't personally tooled around with this a whole lot though. SharkD (talk) 20:01, 28 December 2008 (UTC)
HTML accounts for the Web's content; JavaScript accounts for its behavior. It's important to understand the distinction between the two. SharkD (talk) 19:59, 28 December 2008 (UTC)

Textarea font change

Resolved
Reverted. --brion (talk) 22:01, 24 December 2008 (UTC)

Why has the font used in the edit field changed to monospace? Where did we ask for this?

This is far less readable than the default up til now (in Safari/Mac), especially at this font size. We're editing English text and markup here, not computer language. —Michael Z. 2008-12-24 01:11 z

Gah! You're right. That looks horrible (same configuration as you). EVula // talk // ☯ // 01:30, 24 December 2008 (UTC)
Monospace is bad enough, but the default is now 11-pixel Courier for Safari users. Some diacritics and punctuation are indistinguishable, and many foreign scripts show up in proportional fonts anyway. You couldn't make it much worse if you tried. —Michael Z. 2008-12-24 02:06 z
Put it back the way it was or make it a user preference item, please. Thatcher 02:11, 24 December 2008 (UTC)
I like it, it looks good in my browser. But yes it should be a preference item. davidwr/(talk)/(contribs)/(e-mail) 02:12, 24 December 2008 (UTC)
(ec × 2) According to people I talked to on IRC, this is the result of bug 1941 being "fixed". I added a complaint there (as I, too, prefer multispace fonts and use Safari), but haven't reopened the bug or anything (an active bug would be necessary to get the devs' attention, most likely). I've already added the following to my monobook.css as a patch, which might be useful for others as well:
textarea {font-family:'Lucida Grande', sans-serif;}
This sets textareas to use Lucida Grande if available, defaulting to a sans-serif font if that's not available. I use Lucida Grande because that's the default font for Safari for me. I'd be happy to add this as a Gadget if people are interested. {{Nihiltres|talk|log}} 02:14, 24 December 2008 (UTC)
font-family: inherit; would probably restore the old behaviour in Safari (other browsers override with monospace, perhaps because their programmers thought that everyone would use them for programming).
But this is beside the point. The display is significantly worse for millions of editors who won't edit their CSS. Thanks for the Bugzilla pointer. —Michael Z. 2008-12-24 02:17 z
I'd use inherit, but from what I understand use of the inherit value can be inconsistent across browsers. Since my personal monobook.css will have an effect no matter which browser I use, I prefer code that will give me more or less (depending on what fonts are available) consistent results when I use other computers. {{Nihiltres|talk|log}} 02:40, 24 December 2008 (UTC)
This is a non-issue. "Millions of editors" already see a monospace font when they edit Wikipedia articles, it was the small minority of Safari users which were seeing the non-standard proportional font when editing. As mentioned you can still override it using your userspace CSS. If you wish to push through a change on all editors though, you'll need to start an RFC most likely and then, assuming you find consensus there, open a bugzilla report requesting the change. —Locke Cole • t • c 03:22, 24 December 2008 (UTC)
You've pushed through a change for the worse onto a small minority of Safari users, and you're using “millions” of unaffected editors as an excuse. Can you explain how this is not a load of horse shit? —Michael Z. 2008-12-24 08:51 z
Not all of the "millions" of non-Safari users are unaffected, either: my copy of Opera is using a different monospace font now. --Carnildo (talk) 10:26, 24 December 2008 (UTC)
OTOH, this issue might need further exploration. Perhaps another bugzilla entry with a specific fix for Opera is in order? —Locke Cole • t • c 10:56, 24 December 2008 (UTC)
Consistency across browsers is never a change "for the worse". I'm sorry you dislike the change, but this is the way it's displayed (and has been displayed) in all other browsers (that is, a monospace font, not a proportional font). —Locke Cole • t • c 10:56, 24 December 2008 (UTC)
Your comment (besides missing all the facts of this case) seems to be based on a fundamental misunderstanding of what the web is. It is not for the content provider to decide how the user sees (or hears/feels) the content. It is the user and his user agent that must have absolute control. What the content provider can do is provide the semantic information, that is needed to make optimal decisions. If users are now seeing this differently, it is by their choise of browsers and user agents. Most likely they have made this choise because it maches their preference for look and feel. You need to learn to respect this choise. -- Petri Krohn (talk) 15:11, 24 December 2008 (UTC)
This change effects not only Safari, but also IE6. Version IE7 seems to reject this ugly-ass directive and use the default font anyway. Good for them, but please do not ask me to "upgrade" to IE7. -- Petri Krohn (talk) 15:19, 24 December 2008 (UTC)
STRONGEST POSSIBLE OBJECTION - put it back the way it was, or I will leave Wikipedia for good!!!! -- Petri Krohn (talk) 14:09, 24 December 2008 (UTC)
Petri, I object as strongly, but let's not go over the top. Although this seems to have been imposed from “on high”, there's no reason why the usual mechanisms can't be used to work this out. —Michael Z. 2008-12-24 16:19 z
Good to see that strong language had the needed effect. I must still apologize to User:Locke Cole for my outspokenness. User should give developers credit for their efforts, intead of fighting them on talk pages. -- Petri Krohn (talk) 02:25, 25 December 2008 (UTC)
Strong language did not have any effect. There have been plenty of cases where people used much stronger language than here and were ignored by the devs/sysadmins (e.g., the rollbacker group, where I think some people may have actually left Wikipedia). What had an effect is people raising legitimate objections to the change, which caused it to be reevaluated and reverted on its merits. —Simetrical (talk • contribs) 20:08, 25 December 2008 (UTC)
Comment - I don't use Safari, and I'm accustomed to fixed-width fonts in the text area (it never occurred to me there was an alternative). But I know how annoying it is when something you're used to suddenly changes. This seems to be a fix to solve something that doesn't seem to have been a problem, so I'm inclined to suggest putting it back, at least until a user preference or gadget can be set up to make it user defined. —  Tivedshambo  (t/c) 14:33, 24 December 2008 (UTC)
It's a funny thing. When a checkuser/developer made a change to the checkuser function that would be of great benefit to the checkusers, another developer reverted it because it had not been properly discussed, leaving an insulting reason in the change log [8]. But when a developer thinks the way text is displayed for a significant number of users needs to be changed without discussion, it's "for our own good." Thatcher 15:05, 24 December 2008 (UTC)
According to that developer. Not all other developers were asked, either. I don't think there's any good justification that's been presented yet for changing browser defaults here. I'm not going to revert it, though: developers don't wheel-war. I'll try to get the input of Tim or Brion.

The other case you cited, by the way, is . . . well, completely different. First of all it involved a senior developer reverting a junior developer's change, which doesn't really require justification beyond taste and tends to happen on a daily basis. I count 11 commits reverted by Brion yesterday with various (usually one-line) explanations. And he's not always civil either. Second of all, how long information on users is kept is an issue of major community importance and is dictated in a lot of organizations from the very highest levels, like Google's recent commitment to keep user info for only nine months, and planning to cut that to six. It's not something for an individual developer to decide, let alone a junior one. —Simetrical (talk • contribs) 18:10, 24 December 2008 (UTC)

Welcome to how the software world works. It's much the same in the commercial sector -- except that marketing also gets to put in their two cents. QA folks can sometimes push back on this -- if the area the defect occurs is supported, & it violates the spec or some relevant standard -- but in the end, the programmers call the tune. -- llywrch (talk) 18:14, 24 December 2008 (UTC)

Now this same formatting is occurring on all WMF wikis. This sucks, and is seriously making me reconsider my on-wiki plans for the holiday weekend. EVula // talk // ☯ // 16:25, 24 December 2008 (UTC)

OK, here's a potential solution for the English Wikipedia, at least, pending a reasonable consensus and some testing in other browsers: MediaWiki:Common.css#Can_we_add_.22textarea_.7Bfont-family:inherit.3B.7D.22.2C_please.3F. I'd prefer not to use the inherit value, but as long as it doesn't actually break anything in the other browsers, it would be an acceptable solution—and specifying font-family the way I did above is too crude for a general fix: it's specific to my preference. {{Nihiltres|talk|log}} 20:22, 24 December 2008 (UTC)
Never mind, it alters other browsers to use the inherit value explicitly. {{Nihiltres|talk|log}} 20:31, 24 December 2008 (UTC)

Browser users survey

Let's just find out how this stacks up. Please let us know if you use Safari or another browser affected by the CSS change as your main browser or alongside other browsers, and how this change affects you: positive, neutral, or negative. Let's please try to stick to factual and practical observations. —Michael Z. 2008-12-24 16:50 z

  • Negative—I am a full-time Safari user, editing Wikipedia for 4-1/2 years with it. My edit field has changed to 11px Courier, from 11px Helvetica Lucida Grande. I have previously tried other fonts in my CSS, and found the default to be the best, and a reason to prefer this browser. The change to Courier 11 in Safari is glaring, less readable in my opinion, and makes certain characters which I regularly use completely indistinguishable. The stated benefits are irrelevant: I do a lot of table and template editing, but this is still a small minority of wikitext I deal with, and I do it in both the proportional text field and in a powerful text editor where I make a good monospace font choice (Monaco 12 or Consolas 12). I am also upset with my perception that unaffected users have imposed this change and are telling us to be happy with it. —Michael Z. 2008-12-24 16:50 z
  • Neutral but only because I came here and learned about the monobook fix. I chose Safari over Firefox precisely because I like the look and feel of Safari better, as Petri Krohn put it, forcing me to see things your way negates my freedom of choice to do it my way. Thatcher 16:54, 24 December 2008 (UTC)
  • Negative I'm honestly very pissed off about this change; it affects every single wiki, and makes editing very, very difficult. I've been using Safari as long as it's been out, and I'm not about to switch to Firefox just for wiki stuff. EVula // talk // ☯ // 18:10, 24 December 2008 (UTC)
  • Negative—I use Safari much of the time, and I find it very disruptive to find that the normal editing field has suddenly been squelched into ugly monospace, which offers a significant disadvantage in that en and em dashes (– and —) are hard to distinguish (and I hate using the HTML entities like &mdash;). It's my position that this change makes no sense: even as an editor who has done not-insignificant work with templates and tables, I definitely prefer multispace fonts for editing. Given the following: that a significant number of Safari users dislike the change, that the change affects primarily Safari users, and that it's harder to set the field back to default than to set it to monospace (since the inherit value is quirky in some other browsers), I think that forcing monospace fonts for textareas is a bad thing. It's especially annoying given that it applies to all Wikimedia wikis: if this were just a local thing I'd be willing to accept just using my monobook.css and ignoring the issue—now I'll presumably have to update a number of monobook.css pages. It would be better to allow browser defaults: the extreme minority of people who both a) use Safari and b) actually want a monospace font can use their monobook.css. Ideally, this would be a user preference with four options: browser default, monospace, serif, or sans-serif. {{Nihiltres|talk|log}} 18:36, 24 December 2008 (UTC)
  • Negative: On my installation of Opera for Linux, this has changed the font from the browser's font for multiline text boxes of 16-point Courier New, to the browser's webpage monospace font: 24-point Bitstream Vera Sans Mono. --Carnildo (talk) 00:16, 25 December 2008 (UTC)
  • Closing the survey—Thanks for your feedback, but the point is now moot. The change was rolled back. —Michael Z. 2008-12-25 00:25 z

A couple of remarks

  1. The only thing changed was the software default style rules. The English Wikipedia is not consulted on changes to the default software, because its opinion deserves no more weight than any other user of the software (weighted by size) and there are way too many users to feasibly consult them all. Changes to the software do not, will not, and cannot obtain consensus on enwiki before being made in the general course of things. "Where did we ask for this?" is therefore not a reasonable question. If a change turns out to be unexpectedly unpopular after the fact, developers will consider changing back to the old behavior or sharing workarounds.
  2. In this case, adding the line textarea { font-family: inherit; } to MediaWiki:Common.css should go back to the old behavior. No, this won't work in IE because as usual IE is broken, but the change didn't affect IE anyway as far as I know. Actually, of course, this won't work. There looks to be no standard way to undo the change without special-casing each browser. Gecko has -moz-initial as a valid value, but that's not standard and other rendering engines likely don't support it.
  3. The change passed initial review by Brion, but it remains to be seen whether it will pass further review now that attention has been drawn to it. I and at least one other developer I've talked to don't see why browser defaults should be overridden here. —Simetrical (talk • contribs) 18:19, 24 December 2008 (UTC)
Actually, textarea { font-family: inherit; } breaks FF3 by making it inherit "sans-serif" from the body rule in main.css. Anomie⚔ 19:17, 24 December 2008 (UTC)
Of course, you're right. "inherit" means "use the parent element's style", not "use the browser default style". There's no way to say the latter in standard CSS (at least not in the subset I know about or the subset that's commonly implemented). Yet another reason to revert this change. Which, incidentally, I did for the time being in r45005, after it was pointed out that it caused a regression for Lingala in Safari, but we still have to wait for Tim or Brion a) for this to have any effect, b) for the issue to be decided definitively. (Which makes my revert kind of pointless, I guess.) —Simetrical (talk • contribs) 19:41, 24 December 2008 (UTC)
"initial" is a valid keyword in CSS3, it turns out. But I don't think that does what we want either, actually (nor -moz-initial). It sets the property to its CSS initial value, not to the browser default. CSS mostly considers the browser stylesheet to be just another ordinary stylesheet, and there's no way to say "ignore this last rule and fall back to some other rule" in the cascade. So there's probably no solution at all. —Simetrical (talk • contribs) 20:01, 24 December 2008 (UTC)
I think I'll just go and add textarea {font-family:sans-serif;} in as a Gadget. It's just about the only workable solution I can see that doesn't involve editing one's monobook.css, at least until we can get more developer input. {{Nihiltres|talk|log}} 20:31, 24 December 2008 (UTC)
There, we have a Gadget fix. {{Nihiltres|talk|log}} 20:57, 24 December 2008 (UTC)
It looks like WebKit's (Safari's) default style sheet (referrer) includes textarea { font: -webkit-small-control; }, which resolves to "Lucida Grande" font in my browser. This makes sense, as lucida is about the best broad-coverage Unicode font. —Michael Z. 2008-12-24 21:34 z
The gadget still doesn't make it look like it used to look, but holy shit is it head and shoulders above that horrific monospaced font. They have their place in the world, but the editing field is not one of them... EVula // talk // ☯ // 21:51, 24 December 2008 (UTC)
It might be interesting to try adding in -webkit-small-control as an option before sans-serif in the Gadget. If non-WebKit browsers disregard it (as they should, since it would be invalid for them), then it might make Safari act entirely as it should without screwing up the Gadget for, say, Firefox/IE users who are tired of monospaced font (which is why I used sans-serif in the Gadget and not the inconsistent inherit)… The concern here is that this starts descending back into the realm of interesting hacks, and I'd want people to test not only other browsers, but also other browsers that also use WebKit—I can just imagine it borking in Chrome or something. If people are willing to test it, though … it could be a useful option. {{Nihiltres|talk|log}} 23:55, 24 December 2008 (UTC)
Brion has rolledback the change, so it's effectively not an issue anymore. Phew. EVula // talk // ☯ // 00:18, 25 December 2008 (UTC)
 Well, that resolves things… :) {{Nihiltres|talk|log}} 01:17, 25 December 2008 (UTC)

I think the change to monospace in textareas is a good one. SharkD (talk) 21:06, 29 December 2008 (UTC)

Optimizations not important?

There has been considerable whining[9][10][11] on Bugzilla that optimizations (whether for performance or otherwise) are not important. Are optimizations such an anathema to WikiMedia developers that they need to be quashed on sight? Are developers (who are for these very reasons not likely to contribute anything themselves) so delicate and in need of mollycoddling that even minor syntactical changes are considered a dangerous hazard? Are changes that promote consistent syntax throughout code really so dubious that they must be averted? To me this all seems like nothing but bureaucratic jockeying designed to draw attention to people whose only outstanding qualities are that they have superior SVN commit rights. SharkD (talk) 06:32, 27 December 2008 (UTC)

To quote Knuth: "Premature optimization is the root of all evil." If you can't at least double the speed of an operation that, through profiling, you already know is slow, don't bother. As a rule of thumb, a change in choice of algorithm can provide a worthwhile speedup; tweaking the current algorithm can't. Optimizing code that doesn't need it will just make things more fragile and harder to read, and if the code is being run through an optimizing compiler, hand-optimization can make the code slower by interfering with the optimizer. --Carnildo (talk) 08:53, 27 December 2008 (UTC)
All things JavaScript are known to be slow. SharkD (talk) 10:14, 27 December 2008 (UTC)
So? That doesn't make what I said any less true. --Carnildo (talk) 10:30, 27 December 2008 (UTC)
And what you say does not make anything *I* said any less true. 1) *Everything* JavaScript is slow, and is therefore in automatic need of optimization. 2) It's not a matter of optimizing a particular routine; rather it's one of changing how the code is typed or entered to overcome common inefficiencies introduced by *coders*. 3) I've *already* demonstrated an performance increase by a *factor* of 10-20. 4) Code readability in this case is subjective and based on one's own previous experience; if one familiarizes oneself with proper coding trechniques, then the code is just as "readable"--if not more so--than it was beforehand. 5) JavaScript is not compiled, therefore compiler conflicts are not an issue. Please, enough with the rhetoric and biting. Examine things with a bit more scrutiny before you make blanket statements that have no merit in the situation at hand. SharkD (talk) 05:27, 28 December 2008 (UTC)
The fact remains that you have not shown any perceptible improvement in performance. So what's the point of committing the patch? The only reason to optimize anything is if it makes the user experience better (better perceived responsiveness, etc.), or if it reduces hardware costs (which is not an issue for JavaScript like this, although it comes up sometimes with PHP). —Simetrical (talk • contribs) 16:22, 28 December 2008 (UTC)
This is complete nonsense! I provided a series of test scripts/HTML files when asked that showed a performance increase of 10-20 times—from 14 seconds for sorting a single column down to less than a second. And, I've had to suffer through name-calling and other types of abuse just so you circle-jerks can feel superior for half a day! SharkD (talk) 20:24, 28 December 2008 (UTC)
Sorry, it's been three months, I may have forgotten. I apologize for the misrepresentation, you did say you made a noticeable improvement. But I can't see it. In your test cases, "Optimized" and "Original" both seem to take about as long: 700 to 1500 ms or so. I certainly don't see anything approaching a the order-of-magnitude difference you say you got. (I thought I had already said that somewhere, but apparently my memory is fooling me.) —Simetrical (talk • contribs) 15:44, 29 December 2008 (UTC)
Hmmm. The speedup only seems to occur in IE7. Firefox is (just as) fast in either case. SharkD (talk) 17:49, 29 December 2008 (UTC)
I took some time to look at your patches, as well as the accompanying dialogue. As a former patch contributor who now has commit access, I'd like to give a few tips on making the process less painful:
  1. Don't even try to do code cleanup via patches. It's just not worth the trouble. If you think a particular piece of code is ugly and should be cleaned up, by all means file a bug report saying so, but save yourself and the devs the trouble of writing a patch. It takes more time and effort to review a patch for correctness than it takes to just go over a file and clean it up.
  2. In general, avoid changing code that doesn't need to be changed. Needless stylistic changes make it that much harder to run svn blame on the code to find out who originally wrote which bit and when. There is a balance between maintaining uniform coding styles and avoiding stylistic changes, but, as a patch writer (see the previous point), you're probably best off erring on the conservative side.
  3. Familiarize yourself with the established coding style used for MediaWiki. However much you might disagree with it, it is the style we use for MediaWiki, and if you do make style changes, they're much more likely to be accepted if they bring the code closer to the established style than if they move away from it. Note that the JS files are currently more of a mixed bag than the PHP ones, even though we could (and probably should) apply mostly the same standards to both. Some of the style guidelines followed in most of MediaWiki code include:
    • Indentation is done with tabs (not spaces), one tab per indentation level.
    • Put spaces inside parentheses (like( this ), not like(this)) in all contexts. Also put spaces after commas.
    • Don't put a space before the opening parenthesis after function names nor after loop/conditional keywords like if or while.
    • Use curly braces ({ }) around loop and conditional bodies, even where not necessary. This isn't as strictly followed in my experience, but at least you shouldn't remove them if someone else has used them.
    • Put the opening curly brace on the same line as the loop/conditional keyword, preceded by a space. Put the closing curly brace on a line of its own, unless followed by else / elseif / else if, which go on the same line as the closing brace.
    • Name variables etc. likeThis, not likethis or like_this, at least in PHP. Class names start with a capital letter, pretty much everything else with a lowercase letter. Constants are named LIKE_THIS. There are some prefix conventions followed in PHP code (such as wgFoo for global variables, wfFoo for global utility functions, mFoo for member variables of some (but not all) classes), but no major ones that I know of in JS (except for the wgWhatever JS variables, which loosely correspond to similarly named PHP globals).
    • When in doubt, follow the existing style of the code you're editing.
  4. As you've already been told, the shorter and simpler the patch, the better the chance that someone will take the time to review and commit it. Where possible, one patch should do one thing. In particular, if you find that a piece of code is both broken and ugly, you should ideally write two patches: one the fixes the breakage it with minimal changes, even if that leaves it even uglier, and another that cleans it up. And concentrate mostly on the first one.
  5. Mark your patches clearly. If a bugzilla entry has more than one patch attached, it can be very confusing for someone skimming the report to know just what they're supposed to review and/or apply and against what. This may be a flaw in bugzilla's UI, but you have to live with it. In particular, don't just describe your patches in comments — use verbose descriptions that fully describe each attachement. Remember to mark superseded patches as obsolete. Good descriptions might be:
    • Patch A, fixes scoping bug in someFunction() as described in comments, applies against 1.44.0-wmf.2 (8fd6c9c).
    • Patch B, cleans up convoluted control flow in someFunction() after bug fix, OPTIONAL, apply only AFTER patch A.
    • Patch A2, supersedes patch A, fixes scoping bug in someFunction() as described in comments, applies against 1.44.0-wmf.2 (8fd6c9c).
    • Patch B2, supersedes patch B, cleans up convoluted control flow in someFunction() after bug fix, OPTIONAL, apply only AFTER patch A2.
  6. It's not that hard to get commit access for MediaWiki. Just get a few simple patches accepted, so that people can see that you can write correct code, follow style guidelines and, most of all, work with others. Then ask brion to create an svn account for you.
—Ilmari Karonen (talk) 17:45, 27 December 2008 (UTC)
Your "standards" are for the most part being ignored in the files in question. If I'd noticed any systematic use of a single coding style, I would have abided by it as well. Thus you are setting a double-standard. SharkD (talk) 05:30, 28 December 2008 (UTC)
You missed the point of what he was saying. The fact that there are standards only means that when you are writing new code, it should follow those standards, and that any changes you make to existing code should bring it into closer compliance with those standards, not move further away from it. Most of your patches contained reformatting that moved away from those standards, and reviewers didn't want to spend the time to separate out the functional changes from those issues. Hence the suggestion to break patches down and be more clear about what they're doing. Just because there are standards doesn't mean that anyone needs to go round enforcing them on existing code, or that the fact that no one has done so is an indication of a "double standard" as you claim. Happy‑melon 11:27, 28 December 2008 (UTC)
I must apologize, I missed this distinction the first time I read his post. However, the fact is that I've repeatedly been browbeaten with "standards" 1) without sources deliniating these standards; 2) without clear examples of the standards being followed within the code itself; 3) and with standards that either go against or aren't listed above. This has lead to a lot of anger and consternation on my part. SharkD (talk) 12:26, 28 December 2008 (UTC)
Stylistic standards are covered at mw:Manual:Coding conventions and in docs/design.txt. These pretty much cover only superficial issues to ensure the code looks uniform, which in the end your patches were okay on. There are no written standards on what constitutes good coding style in the somewhat deeper sense of things like "avoid code duplication" and "don't optimize unless you know you're making a difference": those are things that need to be learned by experience and can't be listed out explicitly, because it's not clear without experience how they should apply in any particular case. And yes, they're often subjective, and vary by language and project and person.

It is not expected that nontrivial patches will be accepted on the first round of review, or on the second. As the patch changes and it gets reviewed repeatedly, more issues are likely to crop up with each round of review. This is expected and should not be an issue ― eventually the patch will reach an acceptable form and be committed, unless it's decided that it's not desirable. Partly this is because, in fact, the issues you're being reviewed on are partly subjective, and so you can't get a good feel for them without experience in the project or another project with similar values. Part of the reason people are required to get patches accepted before they get commit access is so that they become familiar with the unstated and subjective preferences of the project's developers and don't annoy people by checking in code all the time that others will think is ugly.

Sometimes it's not clear whether it's desirable (to anyone, including the reviewer) until a substantial amount of work has gone into the patch. This is unfortunate, but it's life. The same thing happens when you're writing code by yourself: you can start on a solution and put a whole bunch of work into it only to realize it's no good. The problem is inevitably amplified when the author and reviewer are different people and may have somewhat different expectations.

Nobody is picking on you. I had to submit patches on Bugzilla to get commit access too. Many of them went unreviewed for months. None was committed for two or three months, IIRC, and some if not most still aren't committed. Many of them had to go through multiple cycles of review. Some of them went through some review and improvement and were only rejected after that. Some were rejected in advance. I submitted patches with screwed-up formatting a lot at first until some developers pointed me to how to do it right (I was actually constructing the patch files by hand at first, not using any diff tool!).

I did not, however, throw tantrums about how the developers were picking on me. I patiently did everything I was told by reviewers. I did waste a lot of effort on patches that never ended up getting committed, but I accepted that. And eventually, having gotten three or four patches committed with maybe one or two dozen still uncommitted, I got commit access for my demonstrated diligence, code quality, and willingness to cooperate with people who knew more about MediaWiki. You are not going to get more patches accepted, or get commit access, by getting angry and accusing developers of "bureaucratic jockeying" when you don't get your way. —Simetrical (talk • contribs) 16:22, 28 December 2008 (UTC)

I thought you said you no longer wanted to deal with me? SharkD (talk) 20:17, 28 December 2008 (UTC)
I don't, but my indignation got the better of me. —Simetrical (talk • contribs) 15:44, 29 December 2008 (UTC)

wiki text

Is there any difference (visible or non-visible) between using

[[User talk:Example|<sup>talk</sup>]]

as opposed to using

<sup>[[User talk:Example|talk]]</sup>

?

If there is, which would you recommend?

Deathgleaner 03:42, 29 December 2008 (UTC)

No, there isn't. Personally, I'd use the second, but that's just because I think it looks cleaner. EVula // talk // ☯ // 03:53, 29 December 2008 (UTC)
There is an non-visible difference in the HTML output, but yes, the actual visible part will be the same. Mr.Z-man 04:19, 29 December 2008 (UTC)
I'd use the second too, it does indeed look cleaner. Gary King (talk) 01:49, 30 December 2008 (UTC)

Force Use of Secure Server

I think it would be pretty easy to write a user script that forced use of Wikimedia's secure server. I for one would find this very useful. Alas, I don't know how. Is anyone interested in using/writing a script like this? Cheers, Jake WartenbergTalk 05:32, 29 December 2008 (UTC)

I've never understood the point of using the secure server for stuff like simple browsing. EVula // talk // ☯ // 06:02, 29 December 2008 (UTC)
If someone is sniffing your network sessions or controlling your upstream router and taking an interest in you, you've got way worse problems than getting your WP login hijacked. Way worse. Franamax (talk) 06:36, 29 December 2008 (UTC)
See Great Firewall of China. --Carnildo (talk) 08:27, 29 December 2008 (UTC)
That's not my issue, Carnildo. I guess I am just security minded. But I did find this, which helps. --Jake WartenbergTalk 19:07, 29 December 2008 (UTC)
Images are still currently loaded from a non-secure server and older user scripts may use an older loading method that loads from the non-SSL site, its not exactly watertight security. Mr.Z-man 20:18, 29 December 2008 (UTC)
bugzilla:16822. --MZMcBride (talk) 20:24, 29 December 2008 (UTC)

Accuracy of category timestamp

Hi,

does anyone have experience as to how accurate the internal timestamps are that MediaWiki generates for category entries, in the categorylinks table? Background: I'm coding a bot that relies on that timestamp field in order to find out, via the revision table, which user added an article to a category (e.g., when they nominated it for deletion). This generally works quite well; however, occasionally, I get some wrong results. For example, it appears as if this edit updated the time stamp for Category:Good article nominees on Talk:Southworth House (Cleveland, Ohio), although it shouldn't have.

Is this a known problem in MediaWiki, or is it normal that these timestamps are "unnecessarily" updated once in a while? Does anyone have experience how reliable the timestamps are, in general? --B. Wolterding (talk) 21:46, 29 December 2008 (UTC)

My best guess, and this really is just a guess, is that the timestamp is updated when the sortkey is changed. And that moving the page merely added the page to the job queue to be reparsed, which had not been done when the next edit was made. Such that as part of the rejigging of the links tables as a result of doing that edit, the category sorkey for the GA cat was updated to reflect the new page name, updating the timestamp too. That's just a guess. Happy‑melon 16:05, 30 December 2008 (UTC)

Creation of a 'disambiguation' namespace

Proposal dropped by Anxietycello. Collapsing discussion to shorten scrolling.  – ukexpat (talk) 18:57, 30 December 2008 (UTC) )

Please consider this thread ended by technical/logical veto, see below.

Would it be possible to create a 'Disambiguation:' namespace, and move all the pages there? Since they're not really articles at all, and shouldn't be in the article namespace. - Anxietycello (talk) 23:13, 29 December 2008 (UTC)

Disambiguation pages contain encyclopedic information; specifically, a list of articles that share a common name. There is no need for a separate namespace for them. EVula // talk // ☯ // 23:41, 29 December 2008 (UTC)

A disambiguation namespace would fix lots of small but irritating problems:

  • The article count is off by however many disambig pages there are.
  • Special:Random shouldn't take users to disambig pages.
  • Special:LonelyPages is useless as it lists mostly disambig pages.
  • Special:WhatLinksHere lists disambig pages, and cannot be told not to.
  • It isn't immediately obvious whether a page is a disambig or an actual article from the title alone.
  • Some pages are titled PageName (disambiguation) but most aren't (should be standardised).
  • The (disambiguation) in brackets is ugly and interferes with the actual disambiguation process: there is a "Newton" in the sense of Newton (unit of force) and Newton (surname), but there is no such thing as Newton (disambiguation) - yet we have an 'article' for all three!
  • Having a unique namespace would allow new users to quickly understand what disambiguation is, and what we use it for.

For these reasons (and I'm sure the community could think of more), I think it's worth creating a new namespace. - Anxietycello (talk) 23:56, 29 December 2008 (UTC)

I would kind of like to see this too. SharkD (talk) 00:58, 30 December 2008 (UTC)
The idea got no support at Wikipedia:Village pump (technical)/Archive 33#Disambiguation: namespace. I oppose a separate namespace for a number of reasons. Your listed problems are either not problems in my view or could be fixed in other ways if it was considered significant, probably with less work than a new namespace. It's deliberate that only some have "(disambiguation)" in the name. See Wikipedia:Disambiguation#Page naming conventions. PrimeHunter (talk) 01:04, 30 December 2008 (UTC)
The vast majority of items you list aren't actually problems. The article count is off? Does anyone care? Special:Random has a 1 in 2 million+ chance of taking someone to a disambig page... which is perfectly fair, considering the fact that disambiguation pages are encyclopedic content. Having "(disambiguation)" at the end of a page's name is hardly the end of the world.
Also keep in mind that disambig pages can help nip content disputes in the bud; the argument about which main topic (the state or the country) should be listed at Georgia would hardly be the best use of our time. :) EVula // talk // ☯ // 01:25, 30 December 2008 (UTC)
The article count is very much cared about, as can be seen by a quick scan of Wikipedia:Announcements. Both times the English Wikipedia has hit a millionth article milestone it was followed by press announcements and was reported by some major media corporations (e.g.). There are currently around 105,000 disambig pages, so 105k in 2.7m would be a more accurate assessment (4% of the article namespace). Seems a lot to say they aren't articles (at least, not according to Wikipedia:What is an article?). Anxietycello (talk) 01:44, 30 December 2008 (UTC)
After looking over the arguments of both sides, I would also oppose a separate disambiguation namespace. Gary King (talk) 01:49, 30 December 2008 (UTC)
Fine, but there's still something that you haven't addressed: disambiguation pages contain encyclopedic information. Why shouldn't they be counted in our article count? EVula // talk // ☯ // 01:56, 30 December 2008 (UTC)
They do contain some encyclopaedic information, yes; that's why they're allowed in the encyclopaedia. But that's not enough to make them articles. An article is a page documenting a defined, specific subject. Images contain often a great deal of encyclopaedic information, as do portals, yet these both have been allowed their own namespaces. Why is that? Anxietycello (talk) 02:40, 30 December 2008 (UTC)
Categories are about as encyclopedic as lists, yet lists appear in the article namespace. SharkD (talk) 02:59, 30 December 2008 (UTC)
Categories merely list links of related content (and sometimes that link is rather tenuous, such as Category:Living people), with a bare minimum of explanation; a disambiguation page explains what each link is, and how it is relevant to the term being disambiguated. They are hardly comparable. You are comparing apples to oranges. ;) EVula // talk // ☯ // 04:15, 30 December 2008 (UTC)
What property of categories, portals and (nav)templates allows them their own namespaces, which disambiguation pages doesn't possess? As far as I can see; none, they're just lists of our articles. Apples and oranges may have their differences, but you're not gonna be able to pass off either as a banana - Anxietycello (talk) 04:43, 30 December 2008 (UTC)
I think the differences are mostly technical in nature, and arose early out of necessity. Futher developments in terms of separation of content haven't caught up yet with the rest of the wiki. SharkD (talk) 04:52, 30 December 2008 (UTC)
The differences: templates are meant to be used on multiple pages on a related topic to allow for easy, standardized navigation between the linked articles; categories are meant to mass-organize loosely related articles under a single banner with a minimum of explanation; portals are the only comparable namespace, meant to guide the reader into a particular concept (a religion, a specific species of animal, a particular vein of pop culture, etc). Disambiguation pages don't serve those purposes, but do allow the reader to find the specific term they were looking for, or to view similarly-named articles with an easy-to-parse description of each item (thereby making it user-friendlier than a category, and more useful than just a template).
As I said, Portals are the closest thing to a comparable concept, but even then, the difference is that everything in a thread is directly related to the overall theme of that particular portal, which a disambiguation page does not. To use a previously cited example, many of the items on Orange share no similarity except for the word itself (as in, the fruit has no direct correlation to the French town, just as the Al Stewart album has no direct correlation to the Polish mobile phone operator). EVula // talk // ☯ // 06:40, 30 December 2008 (UTC)
Very well described. Though, by your definitions of portal and disambig, it would seem portals fit my definition of "article" (a page documenting a defined, specific subject) better than disambiguation pages do. So why were they split, and disambigs shouldn't be? It seems to me, at least, that the main reason you want disambiguation pages to be in the article space is because they're already there (if it ain't broke don't fix it), which I don't think is a much of a valid reason. Anxietycello (talk) 15:55, 30 December 2008 (UTC)
To be honest, I consider your proposal as a "it ain't broke so let's fix it" argument, which is just as invalid. ;)
The biggest difference between disambigs and portals is that the focus of a disambig page is far narrower, related by a very slim connection (it's name, and usually nothing else), whereas a portal is a much more thorough page, specifically guiding the reader among lots of topically-related pages. For a real-world example, compare Portal:Theatre to Theatre (disambiguation). Vijay Tendulkar and Othello have absolutely nothing to do with the word "theatre", just as Theater (warfare) or a Pet Shop Boys song have nothing to do with a venue where performances take place (or, alternatively, the concept of performing).
So, to boil it down, it's the scope and range that are the biggest difference between the two concepts (disambig: scope is name only, range is very narrow; portal: scope is concept only, range is very large). They are both encyclopedic content, yes, but the Portal namespace is more than just a "here's exactly what you asked for" type of page, which is (generally) what an encyclopedia article is. (some of the differences between us and a regular encyclopedia come from the change of medium, so there's only so much direct comparison that can be done) EVula // talk // ☯ // 16:13, 30 December 2008 (UTC)
Sigh. Ok, I retract my proposal. Seems a shame, I still think it would be better if there were a disambig namespace, but I guess it wouldn't be without its faults. Since neither's perfect, I'll stand by the consensus. Thanks for the discussion, guys :) Anxietycello (talk) 17:45, 30 December 2008 (UTC)
Er, correct me if I'm wrong, but disambig pages don't deliver "exactly what you asked for" either. Instead, the result is a situation that is ambiguous, and the user is presented with navigational choices that resolve these ambiguities where the server can't do so alone. SharkD (talk) 20:47, 30 December 2008 (UTC)
Further, let's compare Wikipedia with a hardcover encyclopedia. Printed encyclopedias don't have disambiuation pages. When title ambiguity occurs, readers are directed to the index, which is a completely separate portion of the printed work in content, tone and function. SharkD (talk) 20:50, 30 December 2008 (UTC)

<-- Outdent. All articles are in the namespace, but not everything in the namespace is an article. The classic logical fallacy of the undistributed middle. – ukexpat (talk) 02:53, 30 December 2008 (UTC)

This is a rather silly comment. Namespaces are mutable; therefore there is the potentiality for disambig pages to fall under their own namespace. SharkD (talk) 03:02, 30 December 2008 (UTC)
There's the potentiality for Wikipedia to be replaced with a large mound of carved cheese, too. The fact that potentialities exist does not make them appropriate. I suggest that you think much harder about equal-weight disambiguations, and in what namespace the page containing the disambiguation has to be for such disambiguations to work when readers put names into the search box and hit "Go!", or put article titles on the ends of URLs. I also suggest that you remember the readers. We provide a lot of content in the article namespace that is there for readers so that they can navigate the encyclopaedia, from topic outlines to disambiguations. Uncle G (talk) 15:57, 30 December 2008 (UTC)
There are already cases where entering or appending some text into the Go box or URL lead to pages outside the article namespace. For instance, typing "WikiProject" into the Go box leads to the Wikipedia namespace. SharkD (talk) 20:42, 30 December 2008 (UTC)

Veto on logical grounds. Given that the point of a disambiguation page is that it has the ambiguous title, it would make no sense whatsoever to put them in a separate namespace. --brion (talk) 17:26, 30 December 2008 (UTC)

Disabling indexing of non-content namespaces

There's an ongoing discussion at Wikipedia:Village pump (policy)#NOINDEX of all non-content namespaces regarding disabling indexing of non-content namespaces. Comments would be appreciated. Cheers. --MZMcBride (talk) 00:45, 31 December 2008 (UTC)

Broken Template in "Annapolis"

Resolved
 – ukexpat (talk) 01:30, 1 January 2009 (UTC)

Hi, I hope this is the right place to ask. The template "US state capitals" seems to be broken in Annapolis. Seems not to be a problem of the template because it works i. e. in Montgomery, Alabama, but in Annapolis, you see the source code of the template. I don't know what wrong there, could somebody please see after it? --Schniggendiller (talk) 03:12, 27 December 2008 (UTC) PS: I already tried to bypass browser's cache and also to purge server's cache, with no effect.

{{US state capitals}} was not properly closed. Montgomery probably worked because you had it cached. --—— Gadget850 (Ed) talk - 03:26, 27 December 2008 (UTC)
Okay, so after all it was the template itself. But it's odd that I couldn't see your fixing in Annapolis last night, though I cleared my browser's cache several times. But today, after rebooting my computer, everything is okay. So, thanks for fixing it and have a nice day! --Schniggendiller (talk) 20:37, 27 December 2008 (UTC)
Did you try a server purge? – ukexpat (talk) 22:19, 27 December 2008 (UTC)
Yes, I did. I tried browser cache, then server cache. And again after Gadget850 fixed the template, but my browser still showed my the unfixed version of Annapolis. In the German Wikipedia we would say the hamsters are on strike (german) ;-) (some people believe – in jest – the servers are powered with hamsters in running wheels)
Nevertheless, I wonder why I didn't had the idea to check the source code of the template itself. Maybe I would've noticed the missing closing }}. --Schniggendiller (talk) 01:05, 28 December 2008 (UTC)

British flag template

Resolved
 – ukexpat (talk) 01:30, 1 January 2009 (UTC)

There are several pages (such as List of English flags which reference the external site the Flag Institute. Unfortunately that site indexes the flags numerically and if they add a new flag, all the numbers change. EG http://www.flaginstitute.org/index.php?location=10&flagtype=county&flagid=117 . Is there a way of making a template where we can say {{Flag Institute|Yorkshire}} or something, and have the correct link generated and only have ONE place to edit all the index numbers when they change? If so how would it be done? -- SGBailey (talk) 23:47, 28 December 2008 (UTC)

A simple switch statement should do the trick:
{{#switch: {{{1|}}}
<!-- add the places and links below here -->
 | Yorkshire    = http://www.flaginstitute.org/index.php?location=10&flagtype=county&flagid=117
 | Chiantishire = http://www.flaginstitute.org/index.php?location=10&flagtype=county&flagid=110
<!-- if it's not in the list default to the index -->
 | http://www.flaginstitute.org/index.php?location=10&flagtype=county
}}
This will just output the url, you might want to enclose it in link brackets. — Blue-Haired Lawyer 04:12, 29 December 2008 (UTC)
Template:Flag Institute works a treat. Thanks. -- SGBailey (talk) 10:44, 29 December 2008 (UTC)

Userpage

Resolved
 – User referred to Help Desk. ukexpat (talk) 01:30, 1 January 2009 (UTC)

Can someone make it so that the area between my userboxes is the same solor as the area surrounding it? --Melab±1 ☎ 02:36, 30 December 2008 (UTC)

I think this request would be better off at the help desk. –Juliancolton Tropical Cyclone 02:45, 30 December 2008 (UTC)

Synthesis dispute.

Resolved
 – User directed to appropriate discussions elsewhere. ukexpat (talk) 01:30, 1 January 2009 (UTC)

Hello, I have a question regarding synthesis. If I put two sections of facts which are differently sourced together, does it constitute synthesis? I will provide an example Say I have an article called "Comparing apples and oranges".(of course this is not notable, but bear with this example).

Then I have 2 sentences in the section "color" which are like this:

"Apples are red"(source: www.applecolor.com) "Oranges are orange(source: www.orangecolor.com)

Does this constitute synthesis(the real example is much more complex than that, but you get the point.Teeninvestor (talk) 14:32, 30 December 2008 (UTC)

  • Hypotheticals really don't help, so please explain the underlying reason for your question and let's deal with that. – ukexpat (talk) 15:07, 30 December 2008 (UTC)
  • This is not the right page for content disputes and if you have a specific dispute in mind then it's nearly meaningless to ask in this vague hypothetical way instead of pointing to the actual dispute. The evaluation would depend completely on the context. For example, if the two statements together can easily give a certain impression which neither source gave by itself then you are on dangerous grounds. As an example, "X has property Y" and "Property Y usually goes together with property Z" can give the impression that X probably has property Z even though no source mentioned X and Z at the same time. PrimeHunter (talk) 15:11, 30 December 2008 (UTC)
  • Well, I will give a specific context. I am writing an article "Comparison of Roman and Han Empires" and I have structured several sections like this:

    3 paragraphs comparing both(source compares both).

    paragraph gives details about han empire(DOES not compare both)(source is only about han empire).

    paragraph gives details about Roman empire(DOES not compare both)(source is only about roman empire).

    Because of this, I have been accused of synthesis because supposedly, putting 2 paragraphs of facts next to each other is synthesis.Teeninvestor (talk) 15:16, 30 December 2008 (UTC)

Wikipedia:Articles for deletion/Comparison between Roman and Han Empires is where the real discussion is, everyone. Uncle G (talk) 16:00, 30 December 2008 (UTC) {edit conflict}

Wrong noticeboard, this is for technical issues, bugs, etc. The specific issue is being discussed here: [12] and the general issue of comparison articles here: [13]. As well as on the AfD page. Let's not start another discussion elsewhere please, especially not here. dougweller (talk) 16:04, 30 December 2008 (UTC)

Need an image resize option for Template:Geobox

An image resize option is needed for Template:Geobox; currently users of the template have no way to regulate the size of an image, leading to oversizing. A fix or advice on who to turn to for a fix would be of great help. This option has been provided for in other infobox templates. --Pgagnon999 (talk) 17:24, 31 December 2008 (UTC)

It already has one, the 'image_size' parameter. Algebraist 17:32, 31 December 2008 (UTC)

Bug with CURRENTWEEK ?

Is there a bug with the magic word CURRENTWEEK ? It currently shows as week 1 (in my time zone, at least), instead of 52 or 53. See for example Template:Env.
SyG (talk) 18:51, 31 December 2008 (UTC)

That's correct behaviour, ISO week dates being funny things. Algebraist 19:02, 31 December 2008 (UTC)
"The first week of a year is the week that contains the first Thursday of a year." That must be the part where I just go and hang myself. Thanks for pointing that out to me ! SyG (talk) 19:11, 31 December 2008 (UTC)

Buggered up template: {{orthodoxwiki}}

Resolved
 – ukexpat (talk) 01:30, 1 January 2009 (UTC)

Anyone here know how to do templatey stuff? The "Article name" parameter is meant to take the field and display the text as well as link to it at Orthodoxwiki, iostead it appears as {{subst:PAGENAMEE}}, as shown here, as well as many other articles. Can anybody help? – Toon(talk) 18:58, 31 December 2008 (UTC)

I've made some changes to {{orthodoxwiki source}} and redirected {{orthodoxwiki}} there, although I personally think that the code should be at the latter. I hope my documentation makes sense, do give me a buzz if it doesn't. Happy‑melon 19:09, 31 December 2008 (UTC)
Ah excellent, that was easy! Thanks for taking the time, I'm hopeless with the technical stuff! – Toon(talk) 19:11, 31 December 2008 (UTC)

Self reference

Resolved

I could try it, but I thought I'd ask first. Can the "noinclude" bit of a template use the template that it is part of? -- SGBailey (talk) 20:11, 31 December 2008 (UTC)

Yes. This is very common in documentation. Algebraist 20:13, 31 December 2008 (UTC)
God help us if an <includeanyway> tag is added. — CharlotteWebb 21:46, 31 December 2008 (UTC)
Just in case anyone is reading: The recommended way to do template documentation is by having a transcluded /doc subpage. davidwr/(talk)/(contribs)/(e-mail) 23:32, 31 December 2008 (UTC)
Yes, see Wikipedia:Template documentation. Graham87 06:45, 1 January 2009 (UTC)

commas not followed by a space

Some of the articles have non-standard usage of comma, i.e. comas are not followed by a space, for example: "USA,Mexico,Canada." It would be a good idea to have a script periodically scan all articles and add a space after commas in such cases. Thank you. —Preceding unsigned comment added by 59.183.176.207 (talk) 16:47, 1 January 2009 (UTC)

Let's have a read-only bot do some scans and see if there would be any false positives. davidwr/(talk)/(contribs)/(e-mail) 17:18, 1 January 2009 (UTC)
Only place I can think of where an unspaced comma is used would be in a number, i.e. 5,000. --—— Gadget850 (Ed) talk - 19:44, 1 January 2009 (UTC)
You can also see it in the names of organic compounds, such as 1,1-Dibromoethane. Tra (Talk) 21:55, 1 January 2009 (UTC)
And it could be legitimate style in programming or odd quotations. —Simetrical (talk • contribs) 00:19, 2 January 2009 (UTC)
Possibly examples of poetry as well. davidwr/(talk)/(contribs)/(e-mail) 03:51, 2 January 2009 (UTC)

Automatic Signature in template

Hello, I am creating a user template.

When I add this template to another page, I would like the emplate to automatically sign my name using ~~~~

Thank youtravb (talk) 19:30, 1 January 2009 (UTC)

Add your signature to the template page itself, at the end of the message. Whatever is on the template page will be transcluded or substituted when you use the template--Unpopular Opinion (talk) 19:38, 1 January 2009 (UTC)
You can this using the following syntax: ~~<includeonly>~~</includeonly>. HTH AmiDaniel (talk) 19:40, 1 January 2009 (UTC)
Oh, that is cool, like Template:signature thanks everyone! travb (talk) 20:08, 1 January 2009 (UTC)
That is SO clever, I get how it works. ~<includeonly>~~~</includeonly> would also work, or any combination...travb (talk) 20:09, 1 January 2009 (UTC)
Actually, that one wouldn't. — neuro(talk) 20:18, 1 January 2009 (UTC)
Any combination that doesn't put three or more tildes together will work; anything more and you'll start getting it substitute various bits of your sig... Happy‑melon 22:18, 1 January 2009 (UTC)
Resolved
 – ukexpat (talk) 03:18, 2 January 2009 (UTC)

When I try to add an external link (edit diff) which has an equal sign in it, to a template parameter, this external link does not show up.

For example:

{{User:Inclusionist/irony|http://en.wikipedia.org/w/index.php?title=User_talk:Inclusionist&diff=261294465&oldid=261291639}}

Produces only {{{1}}}, not the link in the parameter.

But {{User:Inclusionist/irony|http://en.wikipedia.org/w/index.php?title}}, which has no equal signs, outputs the http://en.wikipedia.org/w/index.php?title link.

Any suggestions or work arounds?

If there is no technical solution, is there anyway to add a link to edit difference without equal signs in the web page address?

Thank you, travb (talk) 20:16, 1 January 2009 (UTC)

You want {{User:Inclusionist/irony|1=http://en.wikipedia.org/w/index.php?title=User_talk:Inclusionist&diff=261294465&oldid=261291639}}, which defines the value of {{{1}}} to be 'http://en.wikipedia.org/w/index.php?title=User_talk:Inclusionist&diff=261294465&oldid=261291639', while your code defines the value of {{{http://en.wikipedia.org/w/index.php?title}}} to be 'User_talk:Inclusionist&diff=261294465&oldid=261291639', achieving nothing. Algebraist 20:18, 1 January 2009 (UTC)
Algebraist you are awesome, thank you SO MUCH. travb (talk) 20:20, 1 January 2009 (UTC)

Done, but with errors on page

I keep getting these yellow triangles. It's not happening on every Wikipedia page, but it is happening some of the time.Vchimpanzee · talk · contributions · 19:34, 30 December 2008 (UTC)

For example... — Blue-Haired Lawyer 21:13, 30 December 2008 (UTC)
It sounds like a message from your browser and not from Wikipedia. PrimeHunter (talk) 21:19, 30 December 2008 (UTC)
Perhaps they are JavaScript error messages for scripts that aren't working correctly, depending on your browser. Gary King (talk) 21:37, 30 December 2008 (UTC)
That is exactly what they are - IE's ever-so-helpful javascript error messages.  – ukexpat (talk) 21:45, 30 December 2008 (UTC)
IE 4 just doesn't hack it anymore. — Blue-Haired Lawyer 22:00, 30 December 2008 (UTC)
Vchimpanzee, when you next see that message in the status bar, double click the yellow triangle and post here the details given about the error. Also, do you have any gadgets enabled and if so, which ones? (you can find this out by going to Special:Preferences and clicking the Gadgets tab) Tra (Talk) 22:46, 30 December 2008 (UTC)
This has also been happening to me, when I view WP:VG. It doesn't seem to do it anymore. In fact, it was removing a rather large transclusion from a particular page that solved it for me. It was causing CPU issues, would you believe it? Weird. --.:Alex:. 21:48, 31 December 2008 (UTC)

In response to Tra, I got "Unterminated String Constant" and something about line 88 and character 109.Vchimpanzee · talk · contributions · 21:28, 1 January 2009 (UTC) It seems to do this regardless of the Wikipedia page. I only have two examples to work with since I found this answer.Vchimpanzee · talk · contributions · 21:50, 1 January 2009 (UTC)

I can't figure out what's causing this. I did some testing in Internet Explorer 7 and I did get that same error whilst testing some code at one point but on refreshing the page, the error went away. This happened with my monobook.js blanked and replaced with a test to see if the IPV6 related scripts were doing anything. The only gadget I have enabled is the one that hides the fundraiser banner, but that one uses only css so it can't be causing the problem.
What I suspect might be happening is that this problem is being caused by a script that involves generating random numbers, and behaving differently depending on the random number. This is why I suspected the IPV6 script might be responsible, as it's set to run on 1% of all page views. The problem could also be related to a fundraiser script since that does different things on each page view. I'm just guessing here, though. Does anyone else know what is happening here? Tra (Talk) 01:14, 2 January 2009 (UTC)

I block the fundraiser. This time I got line 83 and character 109, but I wasn't signed in at the time.

I suppose it's not a problem, but with email sometimes it won't send if I have that triangle.Vchimpanzee · talk · contributions · 15:11, 2 January 2009 (UTC)

When I saved my last post, it was line 127, character 109.Vchimpanzee · talk · contributions · 15:12, 2 January 2009 (UTC)

Cyrillic characters in italics

Resolved
 – Some fonts have fancy Cyrillic italics which do look different from just slanted letters, and that's ok
Cyrillic italic and non-italics

When Cyrillic words are italicized, the letters appear different. For example, (non-italicized) Russian Ближний Восток or Bulgarian Близкия Изток and (italicized) Russian Ближний Восток or Bulgarian Близкия Изток.

The т character changes into an "m" character. Is that how it's supposed to be? Is it okay to italicize Cyrillic words? --Aude (talk) 18:30, 31 December 2008 (UTC)

When would you want to italicise Cyrillic text? Certainly it should not be done if it changes the meaning of the letters. Happy‑melon 18:33, 31 December 2008 (UTC)
Italicized characters were in the Middle East article, in the section on etymology and translations. I removed the italics due to this issue [14], and don't think they are necessary. But, still wonder why this happens, if it's a bug or italics are supposed to be like that. --Aude (talk) 18:52, 31 December 2008 (UTC)
I guess that's the quirks of Cyrillic script, much like our small "a" in Times New Roman. However I believe those literate in Cyrillic can distinguish italic м and т, which are м and т respectively.
Anyway, we do need Cyrillic guys here. I'm not convincing enough... HУтaяtalk2mecontribs 21:38, 31 December 2008 (UTC)

Are you saying that the letter is supposed to change, or that you're seeing it change and you think that's wrong? I don't see any such letter-changing in your example text above. --brion (talk) 00:13, 1 January 2009 (UTC)

I'm having the same "problem" that Brion is, in as much as not having a problem can be a problem... EVula // talk // ☯ // 07:06, 2 January 2009 (UTC)
What font are you using? That seems to make a difference. --Aude (talk) 07:09, 2 January 2009 (UTC)
I found an image (see right) which shows how the italics are different for certain cyrillic letters. Seems that Arial and Times New Roman handle this correctly, in making the italics different. --Aude (talk) 07:27, 2 January 2009 (UTC)

Yes, handwritten "т" in most versions of Cyrillic script looks like "m" in Latin script, so italic "т" is "т". It's also explained in Te (Cyrillic). --Martynas Patasius (talk) 02:11, 1 January 2009 (UTC)

My point exactly. Even if the text is italicised, it wouldn't affect the readability of the text if you're familiar with the script. That also depends on the fonts installed locally. I guess placing italics in the said article is needed because it's a foreign language in English Wikipedia. So, does that settle your case? HУтaяtalk2mecontribs 19:25, 1 January 2009 (UTC)
Thank you for the explanations. I noticed the difference in italic characters, with the Arial, Times New Roman, and some other commonly used fonts. But, such differences do not appear with some other fonts (e.g. Tahoma, Verdana). Either way, sounds like this is not a bug. In the article, I'm not sure if italics are needed, but may leave them non-italicized for now. That will be a WP:MOS question when the article someday goes through good article and eventually featured article candidacy. --Aude (talk) 06:28, 2 January 2009 (UTC)

adding a cite template inside a template

I'm trying to include a citation inside a template (Template:CDCVesselSanitationScore) and use parameters passed to the template as parameters to the cite template but it's not working the way I'd hoped. Can this be done? Or is my only option to use subst?--Rtphokie (talk) 03:23, 2 January 2009 (UTC)

You can use the {{#tag}} magic word, like this. Anomie⚔ 12:34, 2 January 2009 (UTC)
Thanks for the pointer.--Rtphokie (talk) 20:20, 2 January 2009 (UTC)

Can't edit

Can't edit, won't edit.

I've been trying to copyedit women in Sikhism (from the backlog list), but when I try to preview, nothing happens. I do get an (apparantly empty) "index.php" file downloaded from Wiki, but I am unable to open it, so have no info on contents. I am able to edit other pages, so what could be happening? I'm running Firefox browser (that came with) on Ubuntu 8.04.1 Hope someone can helpdick (talk) 03:30, 2 January 2009 (UTC)

In Special:Preferences under the "Editing" tab, try unchecking "Use external editor by default" and "Use external diff by default" if they are checked. PrimeHunter (talk) 03:40, 2 January 2009 (UTC)
I had a similar problem about 2 days ago, whilst editing my Userpage and a couple of other pages. It seemed only to happen on longer pages. Removing the Userboxes to a subpage sorted out the Userpage problem, and section edits worked fine on the longer pages. -=# Amos E Wolfe talk #=- 23:34, 2 January 2009 (UTC)

Geohack broken

Is it just me, or is the Geohack page broken? When I click on a set of coordinates, I just get a screen full of gibberish, but framed by the GeoHack chrome. It's been like it for a couple of hours; in Firefox and Opera. Andy Mabbett (User:Pigsonthewing); Andy's talk; Andy's edits 16:51, 2 January 2009 (UTC)

There have been issues with the Toolserver. Majorly talk 16:57, 2 January 2009 (UTC)

Blockquotes mess with whitespace

Is there any way of using the blockquote tag such that it doesn't ignore all the whitespace in the source text? I've manually added line breaks where needed, but I'd rather not have to go through all the trouble, as it makes the code uglier. Thanks. SharkD (talk) 17:28, 2 January 2009 (UTC)

T8200. See there for discussion and workarounds. I (or someone) should poke Tim to take a look at that some time, it's a pretty annoying and stupid bug. —Simetrical (talk • contribs) 18:05, 2 January 2009 (UTC)

Straw poll on 'trial' implementation of FlaggedRevisions

The discussion on the implementation of a 'trial' configuration of FlaggedRevisions on en.wiki has now reached the 'straw poll' stage. All editors are invited to read the proposal and discussion and to participate in the straw poll. Happy‑melon 17:52, 2 January 2009 (UTC)

parser #if:

I just revamped Template:Keypress to make it able to display multiple keystrokes (see this for the code). I'm not doing something right. It's still displaying all 10 boxes, even though I used the code {{#if:{{{1}}} | code for the box }}. I reviewed the MediaWiki help page for this and my code conforms to that. Why is the template still doing it? flaminglawyerc 04:08, 3 January 2009 (UTC)

<moved from HD> flaminglawyerc 04:22, 3 January 2009 (UTC)
When you say {{{2}}}, that expands to the value of parameter 2 if its specified and the literal text {{{2}}} if not. To make #if behave as desired, you need to supply an empty default value for the case when the parameter is omitted: {{{2|}}}. Anomie⚔ 04:40, 3 January 2009 (UTC)
Happy day.
Resolved
flaminglawyerc 04:52, 3 January 2009 (UTC)

Please test a page in various browsers

I would appreciate it if people with odd browsers could test the tables in this page. The double-arrows are supposed to line up so they're centered with the borders of the cells in the row above them. (I can upload a screenshot if you don't understand.) I've already tested IE7 and Firefox 3 and have observed the desired behavior. Thanks! SharkD (talk) 23:02, 1 January 2009 (UTC)

Seems to have the desired look on Safari 3.2.1 --TheDJ (talk • contribs) 23:42, 1 January 2009 (UTC)
Looks OK in IE6, but in Opera 9.23, the double-arrows are shifted slightly to the left, so the cell borders above are aligned with the left-hand side of the right-hand arrowhead. -=# Amos E Wolfe talk #=- 00:06, 2 January 2009 (UTC)
Is the problem significant enough to worry about? Will readers have problems getting the idea? SharkD (talk) 01:36, 2 January 2009 (UTC)
Looks fine in Konqueror 3.5.10. Are you aware of sites such as http://browsershots.org/ which will show you screenshots of any URL in dozens of browsers?-Mr Adequate (talk) 01:42, 2 January 2009 (UTC)
  • Konqueror 3.5.9: The arrows don't show up, but the "no such character" boxes are slightly misaligned to the left.
  • Lynx 2.8.6: The arrows are jammed to the left, and have no relation to the cell borders.
  • Links 2.2: The arrows are uniformly shifted one and a half cells to the left.
  • Dillo 2: The arrow alignment bears no relation to the cell borders.
Give me some time, and I can probably dig up some more odd browsers to test it with. --Carnildo (talk) 01:43, 2 January 2009 (UTC)
Yeah, I know about it--I was just lazy (and was hoping noone would complain). It also looks like someone beat me to it. SharkD (talk) 01:50, 2 January 2009 (UTC)
Of the 42 screenshots from browsershots.org that I was actually able to look at (about a half dozen had loading errors, including one instance of someone forgetting to turn off auto-update :) ) only Dillo didn't display things correctly. Add to that the ones that Carnildo listed. Is that good enough for Wikipedia? I'm assuming that most of these browsers are uncommon and intentionally disregard CSS, so very little else on Wikipedia can be expected to work as intended. SharkD (talk) 01:57, 2 January 2009 (UTC)
Links and Lynx are the standard text-mode browsers. Most of Wikipedia is at least usable with them, so whatever you're going to be using the results of this test for, it should degrade gracefully for people using them. --Carnildo (talk) 02:36, 2 January 2009 (UTC)

I am trying to test the existing (i.e. unmodified) table format at browsershots.org. However, today I am receiving an "Browsershots was blocked by en.wikipedia.org/robots.txt" message. This did not occur yesterday. Has robots.txt been modified in the meantime? SharkD (talk) 20:54, 2 January 2009 (UTC)

I tried again tonight and it worked. This represents the current state of the template: User:SharkD/Sandbox/JIT Table Old. Could you please try it in Lynx and Links and compare it to the version above to see if it acts more well-behaved? SharkD (talk) 05:11, 3 January 2009 (UTC)
Works fine in Links, Konqueror, Firefox 2, and Dillo 2; Opera has the empty cells of the bottom two rows too wide, leading to the first and last content cells being misaligned; Lynx makes the cells of the bottom two rows too narrow, leading to misalignment of the later cells. I still can't get Mosaic to display the page rather than download it. --Carnildo (talk) 09:15, 3 January 2009 (UTC)
For some reason the latest version of Dillo that is being tested on Browserhots is 0.8.6. SharkD (talk) 05:04, 4 January 2009 (UTC)

EM or EX in images instead of PX

Stuck
 – Not currently technically feasible

Some images (such as those generated by the {{music}} template) are used as a substitute for text characters. In these cases it is more optimal to use size units that are tied to a given font type, such as EM or EX. I'd like to suggest that these units be supported in images as well. Thanks. SharkD (talk) 20:06, 2 January 2009 (UTC)

Not possible. The server has to generate an image thumbnail with a pixel size. Zetawoof(ζ) 22:15, 2 January 2009 (UTC)
Didn't think of that. However, the image can be rendered in the browser at any size, regardless of its pixel dimensions. SharkD (talk) 23:05, 2 January 2009 (UTC)
The template uses SVG images. I guess once browser support is more widely available, a new and more flexible means of embedding them could be devised, since in these cases pixel dimensions can be safely ignored. SharkD (talk) 23:07, 2 January 2009 (UTC)
T9048. Client-side scaling tends to look terrible. In some browsers it's not so bad, but in some major ones (last I checked) it's nearest-neighbor or some similarly terrible algorithm. MediaWiki therefore avoids all client-side scaling. When we actually serve client-side SVGs, or when browsers reliably use good-looking image scaling algorithms, we could allow this. —Simetrical (talk • contribs) 00:03, 4 January 2009 (UTC)
Yeah. Here's to hoping that the situation will improve at some time in the future. SharkD (talk) 05:02, 4 January 2009 (UTC)

Which database dump includes template pages?

Resolved

I looked at Wikipedia:Database download but I can't find the explicit answer to this question:

  • Which database dump includes template pages, category pages, portal pages?

Thanks. Lightmouse (talk) 21:52, 2 January 2009 (UTC)

, as it says on the database dump download page (link will break when a new one is released), pages-articles.xml.bz2 contains "Articles, templates, image descriptions, and primary meta-pages". Category.sql.gz contains category information, while categorylinks.sql.gz contains info about which articles are in which categories. For portals, you'll have to download pages-meta-current.xml.bz2. Hope this helps. Graham87 10:24, 3 January 2009 (UTC)

That is exactly the information I need. Thank you very much. Lightmouse (talk) 10:28, 3 January 2009 (UTC)

Bug in Special:NewPages?

When viewing from the back of the unpatrolled log at Special:NewPages [15], there are always a few entries missing compared to using the API [16]. I've been trying to figure out why, and I'm at a loss. Is this due to a bug somewhere or am I missing something? Brad 22:59, 2 January 2009 (UTC)

This may be related to a collection of problems: see two threads above. Geometry guy 23:23, 2 January 2009 (UTC)
I doubt it is related. This has been a consistent problem for several months. Brad 23:41, 2 January 2009 (UTC)
NewPages doesn't show redirects by default. Mr.Z-man 05:24, 3 January 2009 (UTC)
They aren't redirects. There is something going on here - there are always about 15-20 articles spanning three or four days that show up in the API but not in NewPages. Take a closer look (bearing in mind the two display in reverse order) and you'll see what I mean. Brad 05:28, 3 January 2009 (UTC)
Hmm, I see what you mean. That's very strange, according to the API, Life in Technicolor II was created on 2008-12-05T16:07:45, which doesn't correspond to anything in the page history, its about 7 minutes away from the closest edit by time, and if you add the "user" parameter, it gives User:I need a name, who didn't edit it until 5 days later. Filed as bugzilla:16877. Mr.Z-man 06:38, 4 January 2009 (UTC)

Deletion creation template on the bottom of user pages

Resolved

There were templates on the bottom of all user pages a couple of days ago, where you can check which articles the editor deleted, which articles the editor created.

  1. What happened to that tag?
  2. Where is it now?
  3. And why was it deleted? travb (talk) 01:45, 3 January 2009 (UTC)
I guess you are thinking about the box at the bottom of user contributions pages. It is still there. PrimeHunter (talk) 02:23, 3 January 2009 (UTC)
Oh Thats it, he he, thanks. travb (talk) 23:58, 3 January 2009 (UTC)

Uploading image from other language wikipedia.

Resolved
 – Save to disk and upload either here or to commons

I wanted to upload an image for the article Kanshi Ram. I did not have the one, but on searching, I found that wikipedia in urdu and hindi language does have this image. I tried to upload that image to english wikipedia, but i could not do that. Will anyone please help me to do this? Ganesh Dhamodkar (Talk) 06:32, 3 January 2009 (UTC)

You have to download it to your computer and then re-upload it to the en. Wikipedia. Or maybe just upload it to Commons instead of doing it on every language Wikipedia. Little Red Riding Hoodtalk 07:47, 3 January 2009 (UTC)
Indeed, if the image is free use (as opposed to Fair Use), uploading it to Commons is the best thing to do. EVula // talk // ☯ // 00:00, 4 January 2009 (UTC)

Problem with alphabetical sorting of importance/class categories

In the new Category:Top-importance_and_Start-Class_MCB_articles page I've created, the member pages are not sorting properly. Instead of alphabetically sorting by the first letter in their mainspace title, the pages are sorting by the first letter of their Talk page title (i.e., they are all listed under 'T'). How could this be fixed, so that categories may be created for multiple assessment criteria (e.g. for categories of two-element, non-redundant combinations of ["Top", "High", "Medium", "Low", "None"] and ["FA", "GA", "B", "C", "Start", "Stub", "List"]). I assume the solution will be involve accessing and/or editing templates, and not adding categories to talk pages as I have been doing. Thanks in advance, Emw2012 (talk) 02:03, 4 January 2009 (UTC)

Many tools not working

Many of the tools at the bottom of a User contributions page do not work (Edit and action count, Contributions summary, Edit summary usage, Files uploaded and Articles created). Is the toolservers database server down? --Apoc2400 (talk) 01:06, 4 January 2009 (UTC)

A search through the toolserver mailing list shows this message reporting one of the systems, yarrow, is down and that two of the database instances are unavailable. Chatter at Wikipedia:Bureaucrats' noticeboard#Toolserver down in turn states that resolution of this problem is on hold till next week when an appropriate person is able to travel to the new data center. --Allen3 talk 01:34, 4 January 2009 (UTC)
For what it's worth, I'm finally able to run SUL reports again. Dunno if that means everything is back, but it's definitely a step up. :) EVula // talk // ☯ // 19:30, 4 January 2009 (UTC)

Tool request: search across citations

I would like to request a tool that makes it possible to do a keyword search across all pages used as references in a paticular article. In theory this seems pretty simple. However, I just tested multiple-page searches in Google (using the "site" prefix) and couldn't get it to work. SharkD (talk) 05:01, 4 January 2009 (UTC)

It would need to access the page text, pick out the references, stip the wikimarkup, do the search, and then work out where the results actually appear in the visible text. It's certainly doable though. Happy‑melon 11:02, 4 January 2009 (UTC)

RfD

Could i point you to Wikipedia talk:Twinkle#Rfd template? Simply south not SS, sorry 17:18, 4 January 2009 (UTC)

Small edits lead to all sorts of strange changes in page

This has happened to me twice in the last month or so but I got a headache just looking at Bug report. So if someone thinks this is important enough to deal with, or I'm just victim of extremely random event twice, do tell.

Is it possible that you're sometimes accidentally clicking "edit" when viewing an old version of the page? This would start editing from that older version, so when you save you change it will include all those differences as well as whatever change you made. Normally this will display a warning at the top of the page "This is an old revision of this page, as edited by" etc in case that's not what you intended to do. --brion (talk) 18:23, 5 January 2009 (UTC)
Resolved
 – ukexpat (talk) 18:13, 5 January 2009 (UTC)

At the top of my Wikipedia page, I can click on "my Userpage, my talk, my preferences, my watchlist, my contributions, log out". Can I add another Wikipedia page to this selection, for example my user subpage? I think I must use monobook.js. --Voletyvole (talk) 22:09, 4 January 2009 (UTC)

I've got a script in my monobook.js file that adds an extra link to User:EVula/admin, which is where I put all my "I need to check or reference these" links. I don't think I was able to get it to link to more than just one page, however. EVula // talk // ☯ // 23:53, 4 January 2009 (UTC)
The easy way is to use addPortletLink. For example, I use the following code:
// link to new pages in sidebar, plus JS and CSS at top
 
addOnloadHook(function(){
    addPortletLink ('p-interaction', '/wiki/Special:NewPages', 'New Pages','n-newpages', 'The list of recently created pages','',document.getElementById('n-help'));
    addPortletLink ('p-personal','/wiki/User:Algebraist/monobook.css','CSS','pt-css','Your custom CSS','',document.getElementById('pt-logout'));
    addPortletLink ('p-personal','/wiki/User:Algebraist/monobook.js','JavaScript','pt-js','Your custom JavaScript','',document.getElementById('pt-css'));
});
Algebraist 00:02, 5 January 2009 (UTC)
Is there a way to attach an accesskey to those? The thing about mine that I really like is that I can just hit control-1 (on a Mac) and voila, I'm there. EVula // talk // ☯ // 01:26, 5 January 2009 (UTC)
The penultimate argument is the accesskey. I don't use them, so it's empty. Algebraist 01:28, 5 January 2009 (UTC)
This is great, thanks a million --Voletyvole (talk) 17:17, 5 January 2009 (UTC)

Image history broken

When I try to view earlier revisions of File:Replace this image male.svg and File:Replace this image female.svg, blank spaces appear in the thumbnails for earlier revisions. When I click on them, I receive a 403 Forbidden from the upload server. I haven't found any other images doing this (but I haven't looked very hard). What's up? This, that and the other [talk] 06:37, 5 January 2009 (UTC)

This is due to an older error mid-to-late last year which deleted a number of image files. (You'll see discussion of it in archives, I'm sure.) Not all images were recovered, particularly in older versions, though some may still be available archived offsite; we have not done a full sweep for old versions yet. --brion (talk) 17:46, 5 January 2009 (UTC)
See Tim Starling's mail. --Amalthea 17:54, 5 January 2009 (UTC)
Just as an FYI, use of those image placeholders now seems to be discouraged, see WP:Centralized discussion/Image placeholders. – ukexpat (talk) 18:12, 5 January 2009 (UTC)

Database replication issues

As stated on the Toolserver mailing list, the database server Yarrow is currently down. ETA for new hardware and the start of the rebuild is one month. The database transfer from Wikimedia should take about one week.

This does not prevent non-database using functions from working, nor does it affect queries that rely solely on the database for the English Wikipedia.

Cross-wiki queries may be affected however, and while we have access to "centralauth_p" (the SUL database), there may be missing or incorrect results from tools that provide information from those projects.

Thank you for your patience. Kylu (talk) 20:05, 5 January 2009 (UTC)

Code update

I've done the weekly code update on the sites; see mw:This week's updates for a quick summary of recent changes. Mostly not too exciting, but some fixes to recent issues and a few bits here and there. We'll start rolling out some fun stuff for testing over the coming weeks... --brion (talk) 02:12, 1 January 2009 (UTC)

A quick follow-up with input box changes. Inputbox can now be used to search page archives, using this code:
<inputbox>
type=fulltext
prefix={{FULLPAGENAME}}/Archive
break=no
searchbuttonlabel=Search archives
</inputbox>
Try it out for this page:
--rainman (talk) 02:27, 1 January 2009 (UTC)

Just a note: It breaks if FULLPAGENAME has a space, users will have to manually type in the page name using underscores, e.g. Talk:Some_page_name. davidwr/(talk)/(contribs)/(e-mail) 02:41, 1 January 2009 (UTC)

Is this working for other people? I'm not getting results. I got excited and set one up at WT:VG, but I can't get it to work for me. Could it be my search settings or something? ~ JohnnyMrNinja 02:52, 1 January 2009 (UTC)
Gah, there is bug because of which it doesn't recognize talk namespaces without the underscore... Will be fixed in next sync, I've fixed the inputbox at WT:VG. --rainman (talk) 03:01, 1 January 2009 (UTC)
You certainly did fix it, thanks! This will be very useful. ~ JohnnyMrNinja 03:02, 1 January 2009 (UTC)
Would {{FULLPAGENAMEE}} work? It uses underscores instead of spaces (among other escapistisms). --Splarka (rant) 11:34, 1 January 2009 (UTC)
The search box is just fantastic. It'll be extremely useful on pages like WP:AN too! --.:Alex:. 17:40, 1 January 2009 (UTC)
I've added it to the WP:AN sidebar. --rainman (talk) 19:14, 1 January 2009 (UTC)

How's this?


~ JohnnyMrNinja 07:02, 2 January 2009 (UTC)

The search results always seem to include "Village pump" which is not very useful... SharkD (talk) 04:54, 3 January 2009 (UTC)
They all contain "prefix:Wikipedia:Village pump", but that's how it searches the village pumps. ~ JohnnyMrNinja 09:36, 3 January 2009 (UTC)
No, I meant that "Wikipedia:Village pump" was being returned and bolded in the actual search results. The problem no longer seems to be occuring, so nevermind. SharkD (talk) 14:47, 6 January 2009 (UTC)

Browser usage per project

Is there any browser usage data for Wikipedia that is on a per-project basis? I'd like to know if certain browsers are used more often on pages that fall under particular WikiProjects. SharkD (talk) 22:58, 1 January 2009 (UTC)

By project do you mean "Wikipedia", "Wikibooks", "Wiktionary" etc or "WikiProject Oregon"? The former can be obtained fairly easily; the latter would be harder but probably not impossible. --brion (talk) 18:18, 5 January 2009 (UTC)
I meant the latter. Thanks. SharkD (talk) 14:44, 6 January 2009 (UTC)

dummy edit js

I have no knowledge of Java. If I did, I could and would figure this out myself. So: Is there a user script (or anything else) that will automatically make a dummy edit to a page? Not like the "undo" feature automatic, I mean like Twinkle automatic. Like you just press a button and it does it. It doesn't even show you an edit screen. flaminglawyerc 03:08, 6 January 2009 (UTC)

Dummy edit or null edit? There is Wikipedia:WikiProject User scripts/Scripts/Add purge to tabs which gives you a link to purge the cache, which depending on what you might want to accomplish will be enough. It reloads the page apparently, but I'd assume that you want to see the result of the purged anyway. --Amalthea 03:16, 6 January 2009 (UTC)
They both seem to encompass what I'm aiming for: an edit that makes no change to the wikitext. Well, I guess that's closer to a null edit. And no, just purging won't work. I'm working with the template-category issue, which is only solved by a null edit or waiting for the server to work it out. flaminglawyerc 03:30, 6 January 2009 (UTC)
User:Splarka/nulledit.js should still work. Try it with importScript('User:Splarka/nulledit.js') in your monobook.js (or copy it there if you wish to modify it or don't trust me, mwahaha). --Splarka (rant) 08:28, 6 January 2009 (UTC)

parameters to make a template display a different section heading depending on which page it is transcluded on?

Have tried to get this to work on my own but haven't had luck so far ...

Template:LGBT rights table Europe is transcluded on both Homosexuality laws of the world and LGBT rights in Europe (so that you only have to edit the template to update both articles).

I have tried to set it where by using {{LGBT rights table Europe |world=yes}} it would display a header of ===Europe=== , and if transcluded using {{LGBT rights table Europe |continent=yes}} it would display a section header of ===Laws by country=== instead. have tried 2 different ways with if, but neither has worked. Am sure I'm doing something wrong (I used code from a navbox , but it didn't translate well to this purpose). As is now, I have it set to just display the "Europe" header always, so that it looks right on Homosexuality laws of the world, but this makes it display Europe as a section header also on LGBT rights in Europe which doesn't work right. Any help (or pointing me to relevent examples) would be greatly appreciated, thx. Outsider80(User0529) (talk) 07:36, 6 January 2009 (UTC)

You're missing a pipe character between {{{parameter|}}} and ===header===. But maybe you want to do something like:
==={{#ifeq:{{{PAGENAME}}}|LGBT rights table Europe|LGBT rights by country|Europe}}===
Which would switch automatically. Or maybe simply:
==={{{header|LGBT rights by country}}}===
Which would take a header name as a parameter. --Splarka (rant) 08:24, 6 January 2009 (UTC)
Resolved
Thanks, Splarka, the extra pipe worked perfect. many many thanks. Outsider80(User0529) (talk) 09:02, 6 January 2009 (UTC)

Article red link - but the article is there!

Resolved
 – ukexpat (talk) 15:03, 6 January 2009 (UTC)

Hello. A few days ago I created the article for a film called Girl No. 217. When you look at the article, it appears that none of the cast have their own article yet, but I created one for the leading actress, Yelena Kuzmina. If you click on her name it takes you to the edit screen, and you can clearly see that the article exists! Why is this happening and how can it be fixed? Lugnuts (talk) 08:02, 6 January 2009 (UTC)

A purge worked. --Splarka (rant) 08:16, 6 January 2009 (UTC)
Yey! Many thanks. Lugnuts (talk) 10:29, 6 January 2009 (UTC)

Sortable tables & Firefox 3.0.5

Resolved
 – Fix update went live

Some WP:FOOTY users have reported problems with sortable tables and FF3.0.5 which is now happening to me too - on trying to sort a list that uses class="wikitable sortable" I am sent back to the top of the page and the element does not sort the column. I've tried this on random lists including List of Sultans of Zanzibar, Test cricket hat-tricks and the originally reported List of one-club men. As a non-IE user - can anyone check whether IE is affected and whether they get the same issue in their Firefox installs? Nanonic (talk) 23:05, 1 January 2009 (UTC)

I have a feeling it's a change in the sortable code, not the browser. I can confirm that Firefox 3.0.5 and Safari 3.2.1 are broken. Calvin 1998 (t·c) 01:05, 2 January 2009 (UTC)
Seems broken in IE7 also. Franamax (talk) 23:33, 1 January 2009 (UTC)
IIRC it's because the buttons are contained in anchor < a > tags, and there's no link inside them. This results in the page being refreshed or the user being shunted to the "_top" of the page. SharkD (talk) 23:35, 1 January 2009 (UTC)
I can't confirm the issue. The List of Sultans of Zanzibar article at least works fine for me in IE7 and FF3. Maybe I need to do a hard refresh or clear my cache first? SharkD (talk) 23:36, 1 January 2009 (UTC)
I have IE7 and it also works fine for me. PrimeHunter (talk) 23:38, 1 January 2009 (UTC)
(e/c) I do not have the same issue on Safari 3.2.1 (Mac OS X). I note differing reports in the original posting as well for Firefox. I'm guessing it might be a conflict between 2 cached javascript files. Try purging your browsercache to see if that solves anything. --TheDJ (talk • contribs) 23:40, 1 January 2009 (UTC)

Now i see this problem as well. "TypeError: Result of expression 'newRows[i][0]' [undefined] is not an object." wikibits.js (line 672). Which corresponds to for (var i = 0; i < newRows.length; i++) { if ((" "+newRows[i][0].className+" ").indexOf(" sortbottom ") == -1) Last change to wikibits.js diff. I'll ask one of the devs to take a look. --TheDJ (talk • contribs) 23:58, 1 January 2009 (UTC)

I reproduced it on the List_of_one-club_men page with Safari 3.2.1 btw. --TheDJ (talk • contribs) 00:05, 2 January 2009 (UTC)
I can't reproduce this. It would seem as though newRows[i] has a non-array element in it somehow, based on TheDJ's reported error. The only way I could possibly see this happening is if staticRows has some bogus content. This might occur if "for (i in staticRows)" is executing loop iterations on i's that don't correspond to actual array elements. I don't know much JavaScript, but I've been told that "for" iterates over object attributes as well as array contents and isn't safe to use for arrays, so I've changed it in r45304 on the off chance that's what's responsible. It would make some sense, since that was a recent change, as TheDJ points out. —Simetrical (talk • contribs) 00:17, 2 January 2009 (UTC)
Can anyone point me to a Bugzilla thread or elsewhere where these changes are described? I've been working on making multi-span row/col headings work in sortable's, so I'd like to catch up on any parallel efforts. Franamax (talk) 00:33, 2 January 2009 (UTC)
It depends on the type of for loop. If it takes the following form it's OK:
for (var i = 0; i < thing.length; i ++)
However, if it takes the following form it can get screwy depending on the type of object:
for (var i in thing)
I never use this latter form in JavaScript. In some other languages it's safe, but not in JS.
Hope that helps. SharkD (talk) 01:30, 2 January 2009 (UTC)
Simetrical already fixed this in SVN. The new code isn't live yet - those affected will likely have to simply wait until the site is updated.  — Mike.lifeguard | @en.wb 05:12, 2 January 2009 (UTC)
Sortable tables are still broken in IE 6 and Firefox/3.0.5. travb (talk) 06:13, 2 January 2009 (UTC)
Is there a target date for this fix going live? -- Tcncv (talk) 06:17, 4 January 2009 (UTC)

You know, this has been broken for 2 days now. When is this going to be fixed? travb (talk) 10:14, 4 January 2009 (UTC)

WT:TW#Table sorting does not work when either Friendly or Twinkle is enabled. --Amalthea 16:55, 4 January 2009 (UTC)
Still doesn't work for me and I don't use either of those tools. Nanonic (talk) 17:10, 4 January 2009 (UTC)

← But you include User:AzaToth/morebits.js in your monobook.js, which causes the problem. To be exact, the follwoing three functions do: 

Array.prototype.uniq = function arrayPrototypeUniq() {
	var result = [];
	for( var i = 0; i < this.length; ++i ) {
		var current = this[i];
		if( result.indexOf( current ) == -1 ) {
			result.push( current );
		}
	}
	return result;
}

Array.prototype.dups = function arrayPrototypeUniq() {
	var uniques = [];
	var result = [];
	for( var i = 0; i < this.length; ++i ) {
		var current = this[i];
		if( uniques.indexOf( current ) == -1 ) {
			uniques.push( current );
		} else {
			result.push( current );
		}
	}
	return result;
}

Array.prototype.chunk = function arrayChunk( size ) {
	if( typeof( size ) != 'number' || size <= 0 ) { // pretty impossible to do anything :)
		return [ this ]; // we return an array consisting of this array.
	}
	var result = [];
	var current;
	for(var i = 0; i < this.length; ++i ) {
		if( i % size == 0 ) { // when 'i' is 0, this is always true, so we start by creating one.
			current = [];
			result.push( current );
		}
		current.push( this[i] );
	}
    return result;
}

If I remove them from morebits.js, I can sort tables again. If I execute either one of them, it stops working. And I have no idea why. --Amalthea 17:23, 4 January 2009 (UTC)

After reading the above again, it makes sense of course, so once the patch goes live it should work again. Until then, turning off anything that uses morebits or similarily extends Array will also help. --Amalthea 17:34, 4 January 2009 (UTC)
I discovered the same thing, just before seeing your message above. I can only speculate that since the offending statement in wikibits is not executed (most tables have no "unsortable" rows), the error is not detected as long as the builtin version of Array is used. However, by adding the extension methods above, Array is treated differently and Firefox apparently does some up-front validation, detects the error and aborts execution. I hate wierd. -- Tcncv (talk) 19:47, 4 January 2009 (UTC)
I'm not sure what you're saying. The offending statement is executed if you add things to the Array prototype. That's the problem. —Simetrical (talk • contribs) 23:18, 4 January 2009 (UTC)
You're right. My "speculation" above was way off. For the benefit of other readers, what was really happening was a result of the for(var i in staticRows) { var row = staticRows[i]; newRows.splice(row[2], 0, row); }, which was supposed to merge staticrows into newrows. That code was also copying the Twinkle-morebits.js created methods (uniq, dups, and chunk) into newrows. Attempts to use these methods a few lines later as if they were arrays triggered the error. If neither Twinkle or Friendly are active, the Array methods do not get created, they don't get copied into newrows, and all's fine. This change corrects the error. -- Tcncv (talk) 00:48, 5 January 2009 (UTC)
Should be able to put this live today, now that it's not the weekend and we're at work. :) (Otherwise it'd wait until the regular update tomorrow.) --brion (talk) 18:16, 5 January 2009 (UTC)
Didn't have time to poke around to it today (yay catching up on year-end bills after holiday ;). It'll go live with the whole Tuesday batch. --brion (talk) 02:12, 6 January 2009 (UTC)
Went live with the rest of the updates last night. --brion (talk) 19:25, 7 January 2009 (UTC)

Current ongoing server problems affecting new messages, edit histories, backlinks, etc.

Resolved

These were all caused by database lag due to a bug in a batch job moving our text storage from one server batch to another on January 2. The job was stopped and the lag issues were resolved at that time. --brion (talk) 18:00, 5 January 2009 (UTC)

You have new messages bar sticking

Resolved

Has something changed regarding the "You have new messages" bar? It seems to be sticking for me and sporadically not going away. If I visit a page for the first time, about half the time, it's showing the bar. If I hit my browser's refresh button (Firefox) it sometimes goes away. My watchlist is continuing to show it, although I can see that it's not just a local browser caching issue because my watchlist is being updated. --B (talk) 16:30, 2 January 2009 (UTC)

Yes, I had the same problem. Majorly talk 16:57, 2 January 2009 (UTC)
Me, too. It's still doing it. - Hordaland (talk) 18:39, 2 January 2009 (UTC)
Same problem, I have IE.--Ipatrol (talk) 21:26, 2 January 2009 (UTC)
Me three (or four). I'm using Firefox 2.0.0.20 with a 1999 vintage iMac and Mac OS 10.3.9. HTH, Her Pegship (tis herself) 21:50, 2 January 2009 (UTC)
Atleast I'm not going crazy. §hep • ¡Talk to me! 21:58, 2 January 2009 (UTC)

You have new messages (last change).

Resolved

An IP left a message on my talk page unsigned. SineBot signed it. Then a user removed the signature and signed it and SineBot re-signed it and removed users. Then the user signed the comment again, and I reverted it. Since SineBot's re-signing on certain pages I get the "You have new messages (last change)." bar and it doesn't matter how many times I click (last change) it doesn't go away. One of those pages is the main page. Help? mynameinc-Review me 18:15, 2 January 2009 (UTC)

As of now-the problem has disappeared. But if it comes back I will post it here. btw if anyone knows what's wrong, post it here for my future use. Thanks mynameinc-Review me 18:17, 2 January 2009 (UTC)
I've noticed similar problems telling me I've got a new message, when infact it's a message that was posted about 1/2 an hour ago (which was signed). Lugnuts (talk) 18:32, 2 January 2009 (UTC)
Most likely related to this issue. --Unpopular Opinion (talk) 18:45, 2 January 2009 (UTC)
I'm also having this problem. The last edit to my talk page was done by BrownBot, who was delivering the December 2008 WikiProject Films Newsletter at 03:49, 2 January 2009 (UTC) and yet I'm still getting the orange bar, even when I purge a page. SchfiftyThree (talk!) 19:24, 2 January 2009 (UTC)
I've noticed similar problems when creating pages under my user namespace. I.e. redlinks linking to such pages are not changing, and sometimes I am being navigated to the generic "start a page here" message when trying to go to the page. SharkD (talk) 20:48, 2 January 2009 (UTC)

Recent edits not showing up in the edit history

Resolved

A few minutes ago, the revision history for Wikipedia:Requests for adminship/Mister Alcohol showed this as the top edit:

  • (cur) (last) 18:05, 2 January 2009 Davidwr (Talk | contribs) m (32,359 bytes) (Reverted edits by Davidwr (talk) to last version by Straight Edge PXK) (rollback | undo)

Yet, when you open the edit, there are two future edits available if you follow the diff tags.

This happened earlier today to me on I think the same page and it happened to another editor on another page as well. davidwr/(talk)/(contribs)/(e-mail) 20:23, 2 January 2009 (UTC)

Edits not showing up in article histories

Resolved

Many of the edits I've made recently are showing up in my contributions list, but not in the article's histories. For example, my edits to Montgomery County High School (Kentucky), and the vandalistic edits which were made prior to that, which I fixed. Little Red Riding Hoodtalk 20:55, 2 January 2009 (UTC)

And now, when I see a newly-created article in Recent changes (such as Jptnow), I'm getting "there is no such article", and no log history to indicate the article was deleted. I have to click on the red link on the "there is no such article" message and go to the article in edit mode, before I can see the article. Little Red Riding Hoodtalk 21:26, 2 January 2009 (UTC)

Half the recent changes disappeared on me when I refreshed, before coming back when I refreshed again. I've also noticed transclusions not updating and have been getting a lot of 404 errors. --.:Alex:. 22:08, 2 January 2009 (UTC)

Problems reported at the Help Desk

Resolved

The following was reported at the Help Desk and has been copied here for inclusion with other similar reported problems.

A comment, not a question. The Wikipedia servers seem to be getting out of sync. Several times in the last 30 minutes I have edited and saved an existing article, only to be told that it doesn't exist. I have had similar problems while patrolling the new page list. After clicking "Mark as patrolled", I get an error message saying I need to select a specific version. In both situations, I try again (sometimes 3-4 times) and it eventually works. I'm going to take a short wikibreak and let the servers rest for now. Truthanado (talk) 16:35, 2 January 2009 (UTC)

Maybe they are still hungover from celebrating the New Year? – ukexpat (talk) 16:42, 2 January 2009 (UTC)
Or maybe they're still lost in the leap second. Truthanado (talk) 17:01, 2 January 2009 (UTC)
Have you tried refreshing the article? Sometimes I start a new article to be told it doesn't exist. ~AH1(TCU) 17:14, 2 January 2009 (UTC)
Yep, I have this happening too. I've refreshed the article and the same thing happens. Has Jimbo got enough donantions to get a server upgrade by now? Lugnuts (talk) 18:57, 2 January 2009 (UTC)
This is really getting frustrating. Articles are randomly there or not there. Changes may take or not. Can someone kick the servers? TN‑X-Man 18:58, 2 January 2009 (UTC)
The problem appears to be ongoing and also affecting "New message" bars and "What links here". However, this is really an issue for WP:VP/T not the helpdesk. Geometry guy 20:28, 2 January 2009 (UTC)

If 2 or more edits are done to my talk page, the "You have new messages" bar is not in the middle (sort of on the bottom of the bar) and the links don't work. It's fine when only 1 edit is made to my talk page. What's the issue? Can someone fix this? Thanks! MathCool10 Sign here! 03:49, 8 January 2009 (UTC)

Template used for CSD#I8 - inappropriate content

Resolved
 – WT:TW#User talk page templates when issuing I8. --Amalthea 19:33, 7 January 2009 (UTC)

Hi,

When tagging an image with an I8 via Twinkle it does two things that aren't great (one is a Twinkle issue, will report it elsewhere)...

  1. Twinkle inserts Template:Firstarticle if the Talk page is blank. This isn't appropriate for an I8 as it includes text like "Unfortunately, one or more of the pages you created, like ***.JPG, may not conform to some of Wikipedia's guidelines for page creation, and may soon be deleted". I think this is a Twinkle issue so will report it on that page.
  2. It then includes Template:Db-nowcommons-notice. Which is fine, except this template includes Template:db-notice which has text about modifying the article to include a hangon notice, and so on.

Basically the whole effect of this is to make it look like the editor has done something wrong as opposed to it being a technical move of a good quality image to Commons.

I tried fixing the Db-nowcommons-notice template, but decided I didn't trust myself to get it right. Can someone more experienced take a look at this please? For an example, see this diff.

Thanks Unusual? Quite TalkQu 17:40, 3 January 2009 (UTC)

CSS to clear stub icons

(This might not be the place to bring this up, my apologies if this is misplaced.)
The stub templates generally include an image located to the left of the text. Some stub templates achieve this by using a table (which i honestly don't like), and some achieve this by floating the image to the left. The following only really applies to templates which use the "floating method" (which I would argue is more correct).

Sometimes articles transclude multiple stub templates, or may possibly contain content following a stub template (Although I cannot think of an example where anything other than another stub is a "problem"). If the CSS rule #stub{overflow:auto;} were added to the stylesheets somewhere, it would prevent the first stub templates left floated "stub icon" from floating to the left of the following stub template icon. Is there a reason, that I am overlooking, for not implementing this? (note: this could also possibly be achieved by adding the previously mentioned rule to class boilerplate or metadata, although I am not aware of the implications of such as I am not aware of all the instances of said classes) Just a thought. Thanks! ./zro (⠠⠵) 18:53, 6 January 2009 (UTC)

Could this be resolved by use of the {{asbox}} meta-template for article stub templates? --—— Gadget850 (Ed) talk - 19:01, 6 January 2009 (UTC)
It looks like that would solve the visual rendering "issue". Should existing stub templates be converted to use {{asbox}} instead? Is there some project to do such? Also, I would note that while this does solve the "visual rendering issue", it does use tables to achieve this stylistic effect. Also, it incorporates the use of <i>...</i> as opposed to <em>...</em>. Both of these choices seem like bad markup to me. Perhaps I'm out of line here though... is there a policy or guideline on markup that I'm not aware of? I would think that we should strive for best practices in semantic markup. ./zro (⠠⠵) 19:25, 6 January 2009 (UTC)
Looking at {{asbox}} a bit closer, it appears that there are issues in the way it handles categories, so it has not been well accepted. I fiddled around trying to convert a project stub template and see some of the problems. It does look like there is some push to fix it. --—— Gadget850 (Ed) talk - 20:21, 7 January 2009 (UTC)

Not entirely sure if this is the best place to put this, but can't really think of anywhere else more appropriate, so here it goes...

I added the move catlinks code to my monobook.css from Help:User styles ages ago in one of my periodic bouts of fairly pointless tinkering, and have always just put up with its annoying habit of spilling over onto article content on articles in several categories. Been trying occasionally over the last few days however to try and so something about it, and yet everything I do either does absolutely nothing or simply causes the entire catlinks box to disappear. I confess that I am not entirely adept at CSS, and so need somebody who has aclue what they're talking about to help me.

In particular:

  1. Nothing I change seems to have any effect on font size. I'd like to make it fairly small text, slightly larger than the size of superscript. Yet no amount of changing font-size, line-height or for that matter the colour has had the slightest bit of difference.
  2. I think I managed to change the width of the box, no idea why that worked and nothing else does
  3. What do I need to do so that the box floats clear of everything else? I tried changing float:right to clear:right, still no luck. Is there anyway to change it so i appears 'above the line' so to speak next to where the featured article star goes for example?

This is what I have so far: 

/* move the catlinks box */
#catlinks {  
  position:absolute;
  z-index:1;
/*  border: 1px solid #aaaaaa; 
  background: #fafaff; */
  right:1em; 
  top:-0.25em;
  width:25em;
  float:right;
  margin: 0.2em;
  padding:0.2em;
}
 
/* format the catlinks itself */
p.catlinks {
  color: #aaaaaa;
  font-family: Verdana,sans-serif;
  font-size:33%;
  line-height: 1.2em;
  text-align:left;
  text-indent:0;
  text-transform: none;
  white-space:normal;
  margin: 0.2em;
}
 
/* format links in the catlinks (as distinguished from ":" and "|") */
p.catlinks a {
  color:#888888;
  font-size:33%;
}

Thanks in advance. Chrism would like to hear from you 19:31, 6 January 2009 (UTC)

  1. class="catlinks" is no longer in a <p>, use .catlinks.
  2. #catlinks still works.
  3. As you are using absolute positioning, you can't have it clear objects that are already there, since it isn't actually there itself. You'd need to give the objects you want pushed down a top margin, or their container a top padding, but since you don't know in advance the size of the catecory list that is a bit hit-and-miss.
Also, // doesn't work in CSS as a comment (as used in your monobook.css), and might be screwing stuff up. This might work better as javascript. (Also, don't go pasting your 'fixed' code here again, just edit the example above or link to a diff on your monobook.css) --Splarka (rant) 21:03, 6 January 2009 (UTC)
Cheers for the help, I've managed to sort out the CSS problems, and now going to have a tinker either attempting to write some javascript or adding some padding to #sitesub and see what I can do. Much obliged - Chrism would like to hear from you 15:27, 7 January 2009 (UTC)

Page tagged 2007 showing in CAT:CSD today?

I stumbled across this page on CSD duty today. It was tagged with {{hangon}} in August of 2007 which puts pages in CAT:CSD as well. Despite being tagged for 1,5 years now, it did only show up in CAT:CSD today. So I wondered why that was the case. Any ideas? Regards SoWhy 09:47, 7 January 2009 (UTC)

Well, if a category is added by template transclusion, as with CAT:CSD in the case of {{hangon}}, the page is only added to the category when recached (Help:Category#Adding_a_category_by_using_a_template) – possibly the task was "lost" from the job queue at the time, due to server trouble or something similar? haz (talk) 11:01, 7 January 2009 (UTC)
I was just gonna say the same. Seems the most likely explanation. --TheDJ (talk • contribs) 11:06, 7 January 2009 (UTC)
The category inclusion was present in the {{hangon}}-template in 2007. Shouldn't it have shown immediately as Help:Category#Adding a category by using a template specifies? I thought recaching only is needed when the category was added to the template after the page was tagged with it? SoWhy 11:13, 7 January 2009 (UTC)
You're right there; I'm surprised no-one's picked me up on that before as I'm sure I've offered that explanation in another context! I'm out of ideas then... I'm still standing by "random glitch", though. Perhaps if you call up the deletion log from around the time that you noticed the oddity, check either side of it for speedy deletions close to that point, then check the deleted history to see if there are any other similar cases? If it's an isolated incident then I guess that a random server quirk is the most likely explanation. haz (talk) 15:26, 7 January 2009 (UTC)

Main namespace discussion

See MediaWiki talk:Nstab-main for discussion about which text to put in the article/page tab. GracenotesT § 15:18, 7 January 2009 (UTC)


URL resulting in no output

Resolved

Why does this URL not work? It used to work before; does anybody know the correct syntax?

And also, just for a second attempt: does somebody know the answer to this? -- Mentisock 17:46, 7 January 2009 (UTC)

In the first URL, the offset parameter is supposed to have the date that you want to offset from, but 10000 isn't a date. What did the URL show before? Gary King (talk) 18:13, 7 January 2009 (UTC)
At some point in the past, the offset parameter was just a numeric offset, in number of results -- so offset=10000 would skip the 10,000 newest pages and show the next 500 "new" pages after those. Since this isn't efficient to do for large offsets, it was changed some time ago to a timestamp index which can very efficiently start from any arbitrary time in the past.
This does mean there's not a clear way to ask for "the 10,000th and beyond new pages", but you can specify a particular date and time if you like to jump to a point in the past. (You could generate a URL using {{#time}} perhaps.) --brion (talk) 19:13, 7 January 2009 (UTC)
Retrieved from "https://en.wikipedia.org/w/index.php?title=Wikipedia:Village_pump_(technical)/Archive_53&oldid=1148698875"