Wikipedia:Village pump (technical)/Archive AF

Wiki is too difficult for casual users

I have recently made some edits and improvement to the "US Declaration of Independence" article that I think have been well received. As I monitor the article over the past few weeks I have become aware of what a pain vandalism is. (This is obvious vandalism, vulgarity mostly, but also sneakily changing dates to nonsense, inserting persons own names, etc.)

However, my main complaint is how difficult it is for casual users to make contributions to the articles. Wiki seems to rely on way to many details that users (such as editors) can't possibly know. I think more of the process needs to be automated. For example you rely on users to indicate if something is a "minor edit" with an "m". Or "reverted", etc. Why not make those MANDATORY items that must be selected when doing an edit, and the page won't be accepted until they, and the summary field, has been filled out? Further you rely on way too many "tags" in your own little language. Just what the world need: more bored computer science people inventing their own, "perfect" language to end all languages!

Also, I find the editor bizarre. It's the worse of both HTML and NOT-HTML. It's great the you can select a word, and make it ITALICS (for example), but to translate this into a bunch of quote marks is an editing nightmare after that. Frankly, I think all that stuff should be hidden from the user. Something that is bolded and italicized is just a mess of serial quote marks....six I think. I guess that I am off in another direction on your editing philosophy since I don't understand it.

Further I find the way that users document and have "talk" inconsistent. Someone unknowingly reverted my edits, I went round and round just trying to communicate to him that I thought this was wrong. Finally I stumbled onto it (but have been unable to duplicate my "stumble" since then.) Later, I found that he HAD admitted his error, but it it a nightmare to use.

So, in summary, I find the whole "back office" portion of Wikipedia an arcane mess. I realize this is a volunteer effort, so you should be applauded for that. But I think (while keeping the article content in tact) you need to re-evaluate how *casual* editors use the system, and make that process easier. There is a wealth of knowledge out there that I fear is not being shared, because the editing/revision process is so difficult. Improve that, and in turn, you'll improve the whole process.

p.s. It's not even obvious why the "village pump" is the place to go to post this stuff. NateX 23:56, 1 December 2006 (UTC)

For what it's worth, there are a number of people, including Jimbo, who agree with you, and who are pushing for a WYSIWYG way of editing Wikipedia, called Wikiwyg. [1] You can try a demo of it here. It converts to and from wikitext, so it may even be more comfortable for you to edit using that now. I don't know what the plans are for it to become integrated with Wikipedia's software though. --Interiot 00:07, 2 December 2006 (UTC)

Referencing articles

I'd like to use <ref name="lastname-year-firstword">{{cite ...}}</ref> for my entire bibliography at the top of an article, and then use <ref name="lastname-year-firstword"/> as I need to throughout the article, and finally have <references/> include everything that I have at the top (it would be nice if it was everything that I cited but it's okay if not). Is there a way I can hide such a section from the top of an article? --Chris Pickett 23:10, 1 December 2006 (UTC)

I discovered a way to hide such a section using span tags, and someone informed me that it was bad to do so. I wrote a how-to without the span tags (i.e. no hidden section) on Wikipedia:Footnotes. --Chris Pickett 02:46, 2 December 2006 (UTC)

[[MemoryAlpha:articlename]] links: how, when, why?

Randomly browsing WP I came across a large amount of links to http://memory-alpha.org/ and started investigating through Special:Linksearch, but then I noticed that MA used another way to put external links to them: a template, {{memoryalpha}}. That template uses some very strange link // Duccio 20:08, 6 December 2006 (UTC)syntax: [[MemoryAlpha:articlename]]: what is that, when was it created and, most important, why does MA have their own syntax for external links? And how do I know how many times that [[MemoryAlpha:articlename]] has been used, and in which articles? Which is, exactly, the scope of this? There are other external links templates, but this is unique, not showing the well-known external links "double arrow".
Thanks in advance. // Duccio 15:59, 1 December 2006 (UTC)

These are interwiki links; you can find the complete list of prefixes at m:Interwiki map. It's not just Memory Alpha, you can also link google:Wikipedia, etc. Tizio 16:06, 1 December 2006 (UTC)
These interwiki links look very dangerous to me, unless someone comes here with a way to detect them: Special:Linksearch doesn't show them. And what advantages do we get using interwiki links to justify this huge discomfort? Are interwiki links somehow better? // Duccio 16:33, 1 December 2006 (UTC)
"Dangerous" in what way? --TheParanoidOne 22:19, 1 December 2006 (UTC)
The danger lies in their misuse - the redirection of users to other sites without realizing it, in particular when they would have been better-served by a link to another Wikipedia article. They look so like other links that users may not know the difference. On the other hand, they are often used in an appropriate way, to link to articles that are actively not wanted on Wikipedia, in order to give users information without requiring or encouraging an article on Wikipedia. For example, they are used a lot (with a warning) at furry convention to link to the smaller conventions, and I think this is an appropriate use. Wikipedia actively doesn't want articles on small fan events, just as they don't want entries on every bit-player in the Star Trek universe. Instead, users are directed to the appropriate source of information - a wiki specific to that topic. GreenReaper 08:03, 2 December 2006 (UTC)
The danger doesn't really lie in their misuse, as any feature can be misused — it lies in the impossibility to control the misuse. External links can be easily monitored through Special:Linksearch, making it a powerful tool for fighting spam. So again, do we have a tool like Special:Linksearch for interwiki links? If not, do we think we don't need it because those sites are "trusted"? But who decides a website is trusted and doesn't need to be checked agains linkspam? See m:Talk:Interwiki map: I don't see many polls there. One more point: what advantages do we have so that this big limitation is of no importance? What is the problem with making it clear to the average user that a link will send him to an external, non wikimediafoundation-owned website? // Duccio 11:19, 2 December 2006 (UTC)
Interwiki links are a different colour from standard internal links, so they don't look exactly the same. --TheParanoidOne 11:33, 2 December 2006 (UTC)
But they look exactly like a link to wiktionary or wikiquote, while driving the user to a completely external website. // Duccio 12:34, 2 December 2006 (UTC)
But then, we've never really restricted usage of interwiki links to Wikimedia sites. See the Interwiki map for a list of interwikis currently recognized by Wikipedia's wiki engine. Titoxd(?!?) 08:32, 3 December 2006 (UTC)
Indeed. And it's wrong, in my opinion, for the reasons I explained above. (since nobody answered the question if there's a tool like Special:Linksearch for interwiki links, I suppose there's not) // Duccio 11:11, 3 December 2006 (UTC)
But then, we can't do anything about it. The Interwiki map is controlled by meta administrators, not by the English Wikipedia; and to a greater extent, the interwiki map is controlled by the Developers. So, you can try bringing up the case to them, although I do not think you have a good chance of actually persuading them. Titoxd(?!?) 17:10, 6 December 2006 (UTC)
Well sure, I know that now — I didn't when I came here, as you can understand reading my first comment. I am actually not complaining but trying to take it up with everyone else here, and coming here was in fact a good idea since at first I thought interwiki links were limited to wikimedia projects and now I have a better understanding of the issue. Perhaps as you suggest I should open a discussion on meta since this is a wikimedia-wide topic, although it seems that I am the only one who perceives this whole thing as wrong. // Duccio 20:08, 6 December 2006 (UTC)
Considering the fact that such interwiki prefixes can't be created by just anyone, I fail to see what the problem is. It would be nice if the Linksearch page could check for interwiki links, but I still don't think this is some bubbling cauldron of potential abuse. EVula // talk // // 01:44, 6 December 2006 (UTC)
Well, abuse doesn't come in the moment of interwiki prefix creation, it comes when those are misused — and anyone can misuse them, since WP doesn't provide a tool like Linksearch for them. This, in my opinion, makes interwiki links a source of potential abuse.
I am open to any kind of compromise and I'd welcome a feature that takes me away the possibility of checking external links usage in exchange for something, but interwiki links do not add anything to the editor or to the reader, they just have the problems I've been trying to outline here without giving any benefit or advantage. // Duccio 12:36, 6 December 2006 (UTC)
Included in the general "interwiki links" category are links to Wikipedia articles in other languages; for today's Featured Article, that would be stuff like de:Mount Rushmore National Memorial, fi:Mount Rushmore National Memorial, and tr:Rushmore Dağı.
I'm trying not to ask "what's the worst that could happen?", but seriously, when it comes to all the different ways to abuse the link system, bogus interwiki links are pretty low on the list. EVula // talk // // 16:58, 6 December 2006 (UTC)

Firefox memory leak on some Wikipedia pages?

I've run into an odd problem. When I navigate to some Wikipedia pages, Firefox seems to go into some sort of mode that chews up processor time and memory. If I go to Comparison of web browsers, I find that my scrolling goes to hell. Task manager tells me Firefox is using ~50% of the processor, and is leaking memory. Note this is after the page load completes, and it continues, apparently indefinitely. In fact, it continues even when that page is not my active Firefox tab! It's not unique to that page...the same symptoms happened to me last night, and I didn't go to that page (sorry, I don't know what other page/s triggered it last night). I don't know if it's unique to my setup, or something more general. I'm running Firefox 2.0 under WinXP SP2. Can anyone confirm/deny the behavior? –RHolton– 13:57, 1 December 2006 (UTC)

Update: This seems to be related to a Firefox add-in (Linkification, in "thorough mode"). I'll update again if I find out more. –RHolton– 14:17, 1 December 2006 (UTC)

Link tags

I think it would be neat if Wikipedia's software added more of the LINK tags to the header of the site's pages, so that users of browsers that support it (e.g., SeaMonkey, Opera, and, with the proper add-on, Firefox) had some navigation possibilities. As it is, there is one for "Copyright" (which shows up in the "More" menu in SeaMonkey), and also one for "Search" (but that goes to some XML thing, not an HTML page). "Top" would be an obvious addition, and if some of the "Up", "Next", and "Previous" values could be filled in for particular pages based on such things as category structure and succession boxes that could be useful. As an obvious case, pages on years, decades, centuries, etc. should have "Next" and "Previous" go to the succeeding and preceding ones. Who knows; if such a high-profile site as this uses these tags, it might even encourage browser developers to actually support them. (It's been a chicken-or-egg thing so far, with browser makers not doing it because few sites use them, and site developers not doing it because few browsers support them.) *Dan T.* 13:08, 1 December 2006 (UTC)

"Up" would be useful, if hand-added to articles (eg. #PARENT [[Wikipedia:Village pump]]) If it were actually widespread, it seems like the sort of thing that might fit in on the CD/DVD version. Some articles have multiple succession boxes, most have none. Alternatively, previous/next could be similar to Special:Allpages... I don't know how useful that would be, but it would be nostalgic at least. --Interiot 20:32, 1 December 2006 (UTC)

Number of new articles

Is there a way to see how many new articles I've started? It's possible to search by user name on Special:Newpages, but that list only goes back one month. Punkmorten 11:18, 1 December 2006 (UTC)

This would be a good idea; unfortunately, it appears it would be too expensive to do using the current database schema (see bugzilla:4150). Tizio 16:36, 1 December 2006 (UTC)
Thanks for answering. I quote: "That would be one bloody nasty query. I pity the bugger who writes that." So I guess no. Punkmorten 21:22, 1 December 2006 (UTC)
user:Interiot will run such queries on request (against an offline copy of the database). -- Rick Block (talk) 23:33, 1 December 2006 (UTC)
I can't run the query currently, since the toolserver has no data for en.wikipedia.org at the moment. I can run it for other wikimedia sites though, and can run it once/if the toolserver starts working again. --Interiot 23:46, 1 December 2006 (UTC)

Is it possible to tell the number of users watching a given page?

A question forwarded from the Chinese Wikipedia: is it possible to tell the number of users who have a given page on their watchlist? This may be used as a gauge for article credibility. Might such a feature be added in future versions of Mediawiki?

Thanks.

-- ran (talk) 03:49, 1 December 2006 (UTC)

Not a good idea to publish which articles have few people watching them as they would just be vandal targets, SqueakBox 03:51, 1 December 2006 (UTC)
I think there's a feature in a MediaWiki extension to do this, but it's turned off for the reasons SqueakBox suggests. Admins have a list of unwatched pages they can access, but it's turned off for non-admin editors. --ais523 08:52, 1 December 2006 (UTC)
The info action can be used for that, but it's disabled by default for performance reasons (if enabled, you could use //en.wikipedia.org/w/index.php?title=Main_Page&action=info; see MW:Manual:$wgAllowPageInfo for some more details. Tizio 16:12, 1 December 2006 (UTC)

What links here oddness

When I checked What links here for a recently created article David Stuart, I saw there were dozens of talk pages listed, including many about Mesoamerican topics. When I look at those talk pages, I cannot find any link to David Stuart. What is causing this? olderwiser 02:11, 1 December 2006 (UTC)

It's from the embedded template: Template:WP Mesoamerica tasklist. —Pengo talk · contribs 02:17, 1 December 2006 (UTC)
Ah, thanks. I didn't even notice that the to do list was hidden on the template. IMO, just another reason to figure out a way to prevent (or at least moderate) whether links on transcluded templates show in 1) What links here and 2) Related changes. Having to sift through the seemingly ever-increasing multitude of links from these templates greatly reduces the utility of What links here and 2) Related changes. olderwiser 03:05, 1 December 2006 (UTC)

Searching through contribs

I have 40 pages of contribs (20,000 edits) and am unable to search through them by date, user or anything else because I cant get all the edits up on 1 page and use a search button. Searching through 40 pages is not a practical option. Is there any chance of getting an "all contribs" page for each user. It would help tremendously, SqueakBox 01:41, 1 December 2006 (UTC)

On one page, no. You can get them on 4 by sorting them every 5,000 edits, by clicking here. Titoxd(?!?) 01:45, 1 December 2006 (UTC)

Well thanks. Is there any reason why you can get them all on one page or is it just because most people dont use their search buttons and nobody has asked before? SqueakBox 01:47, 1 December 2006 (UTC)

monobook not working?

I have a few things in my monobook that I've copied from here and there that are supposed to customize one's Wikipedia experience somehow. They used to work just the way they were supposed to. The last few days they don't seem to be working at all. Am I the only one with this problem? ~ ONUnicorn(Talk|Contribs) 21:21, 30 November 2006 (UTC)

There have been some changes to the global MediaWiki:Monobook.js (e.g. it is now at MediaWiki:Common.js and some other changes). Nothing that should break user's local monobook.js (yet), though. —Ruud 21:27, 30 November 2006 (UTC)

FA and GA symbols

Why is it that FA are marked appropriately in the top right hand corner but GA are not?Buc 17:53, 31 December 2006 (UTC)

Two reasons: 1) GAs are not "official", and 2) templates that put things in absolute places (like the top right hand corner) are controversial (Template:Featured article is one of the very few templates that does this). There was a template for GAs that was deleted, see discussion about this at Wikipedia:Templates for deletion/Log/2006 March 25. -- Rick Block (talk) 19:53, 31 December 2006 (UTC)

Oh right. Well I was thinking of creating just such a template. I take it there's no point? Buc 20:46, 31 December 2006 (UTC)

Please don't. Anyone can pass an article GA, yet it takes a darn committee to remove then - GA means very little, particularly in relation to the process an FA goes through. SandyGeorgia (Talk) 21:17, 31 December 2006 (UTC)

TOCleft guidelines

Hello. Could I safely revert T-Square_(software) back to TOCleft until this question is resolved? I thought TOCleft was fine and put more information above the fold, and used it in Alexander Kronrod but don't want to cause problems. Thank you in advance. -Susanlesch 04:07, 31 December 2006 (UTC)

  • For reference this is the guideline Google found. Is that one right? If so I see no reason not to revert. -Susanlesch 06:41, 31 December 2006 (UTC)
TOCleft as a tendency to totally misformat an article depending on computer your using to view the article. It looks quite horrible over here, so using TOCleft is strongly discouraged. —Ruud 13:18, 31 December 2006 (UTC)
  • Could you please give me links to 1) "over here" and to, 2) where it is "strongly discouraged"? Unfortunately I have more than one other article whose will need to be rethought. As a former boss who was a graphic designer would say, "change one thing and everything changes." Thanks a lot for your reply. -Susanlesch 17:27, 31 December 2006 (UTC)
  • That's OK. I will change them all. Sad to see the center of everyone's screen used for blank space (requiring more scrolling than might be necessary, and in pages with intentionally short headings to accommodate TOCleft) but so be it. Thanks for your time.-Susanlesch 18:11, 31 December 2006 (UTC)

-Susanlesch 18:35, 31 December 2006 (UTC)

staying logged on

For months my account would stay logged on unless I deliberately clicked on "log out", even when I turned off my computer - that is how I want it to work. However, suddenly in the last few days without my clicking log out, my account is inexplicably logged out, sometimes without my even having turned off my computer or closed my browser (I usually use Firefox), and I have to log on again. Several times it happened while I was in process of making an edit and left the computer for a while, and when I returned and completed the edit and signed and saved it, the edit was signed by my IP address, not my signature, which is aggravating. Has something changed here? Is there a setting that I inadvertently deactivated? Is anyone else having the problem? If it's a Firefox setting, would appreciate hearing how to fix it. (And is there any way to get those IP address edits changed to show my account name?) Thanks. Tvoz 07:35, 26 December 2006 (UTC)

Hmm, just went to look at my preferences, and the "Remember my login on this computer" box and the "send me copies of emails I send to other users" had previously been checked but were now unchecked - I assume the first is causing this problem, but wonder how it could have gotten unchecked by itself - anyone? Tvoz 07:42, 26 December 2006 (UTC)
Just happened to me. Was happily looking at my watch list items and clicked on one of them when I noticed on the page I went to the 'contribute' banner had appeared at the top. Puzzled as I had already 'dismissed' that and found I was no longer logged in. Seems to happen once every few weeks, strange. Most confusing as you don't realise you are logged out until you do an edit or something. Using Opera 9.10 here. Dsergeant 16:06, 31 December 2006 (UTC)
It's because the cookie used to identify you as logged on is made to expire after, I think, 30 days from when you log on so when it expires, you are automatically logged out. Tra (Talk) 16:44, 31 December 2006 (UTC)
Yes thanks. I see all my Wikipedia cookies are now set to expire on 30th January... Dsergeant 16:49, 31 December 2006 (UTC)

New technical feature

I was thinking that it would be very exciting if we had a "hide autoconfirmed users" button at Special:Newpages. Most of our A7 G11 G10 business comes from brand new users, so this would be a really neat way of identifying that. What do people think? Can anyone help me open a bug for this, if you think it's sensibel and feasible? - crz crztalk 14:44, 31 December 2006 (UTC)

Have you seen [2]? That all lists changes for new users. Since we have that page, I would think the developers could do a similar page for new pages relatively easily if they wanted. In the meantime, you can use Special:Contributions/newbies & search for "Created page with" in the summaries to find new page creations. -- JLaTondre 14:56, 31 December 2006 (UTC)

New account problem.

I was directed here from the IRC channel after unsuccesfully trying to resolve this issue there.

I signed up for an account a couple of days ago, but something in the signup process appears to have caused the account to become messed up. I'd prefer to not just create a new account, because it took me quite some time to find a username which wasn't taken or was too close to an existing name without tacking on random numbers.

The process went as follows:

I created the account name 'MrJamesGurney' (with case as specified). I received an address confirmation email, and clicked on the link and was told that my email had been validated.

Now, when I try to login, I'm told the password is incorrect. It's conceivable that I typed my password wrong twice in a row, but it's not something I've done in a very long time. Regardless, the password is supposedly wrong, so instead, I click the 'E-mail new password' button. Whereupon, I'm told "There is no e-mail address recorded for user".. which can't be true, because I've received the email confirmation email and clicked the verify link and been told my address has been verified.

So, obviously something has become messed up with the account, and rather than simply create a new one, is there some way the status can be reset so that I can actually use this account?

I can, if necessary, forward the confirmation email with full headers, to indicate the connection between the username and my email address. —The preceding unsigned comment was added by 67.161.116.15 (talk) 10:14, 31 December 2006 (UTC).

I think you need to check a box in your preferences to actually enable your email after you have confirmed it. You need to contact a developer in this case (try User:Brion VIBBER, User:Tim Starling or file a bug report at http://bugs.wikipedia.org). It might be faster to create a new account, though (User:JamesGurney, User:James Gurney and User:Mr. James Gurney are still available.) —Ruud 13:15, 31 December 2006 (UTC)
Your user entry has been corrupted due to an unidentified MediaWiki bug. Please tell us your email address so that we can mail you a new password. -- Tim Starling 14:51, 31 December 2006 (UTC)

Can we make multiple columns (like in a newspaper?)

On some pages (like Category:Merge by month) the text sorts into 3 columns. I'd like to be able to do this on some articles that have longs lists of short entries. As an example check out Androgyny#Famous_androgynes. Currently anytime anyone adds or deletes a name they have to manually reset the table. Is this possible? Tocharianne 04:42, 31 December 2006 (UTC)

Oh got it! I was following the link above to the discussion on capping overly long signatures (ultimately to this page) and they had a multi-column with the code. (For some reason the Category:merge by month doesn't have any code that would explain the format.)
{{MultiCol}} {{ColBreak}} {{EndMultiCol}}
Nope, it has the same problem as a table--it will still need to be manually updated after any change. What I want is to be able to specify just the number of columns and then have the items be split that way. Tocharianne 05:04, 31 December 2006 (UTC)
It is possible to do with CSS styling rules, but it only works in some browsers. In browsers that don't support it (which includes IE6 at least, which is the largest segment of our readers), it will fall back to a single column. The way to do this is to put <div style="column-count: 2"> ... </div> around the list. I wouldn't recommend it. Mike Dillon 05:09, 31 December 2006 (UTC)
(edit conflict) I was going to mention <div style="-moz-column-count:3; column-count:3;">[insert text here]</div>. This should be what you want, but be aware that this does not work on some major browsers. Among "table-style" templates, there is also {{columns-start}} (with num=2 or 3 as parameter), {{column}}, {{columns-end}} - this set of templates won't make ultra-narrow columns if the page is too narrow - they will drop back to two or one column. Gimmetrow 05:14, 31 December 2006 (UTC)
  • Didn't IE7 break column? Or was that template updated since 6 December? That would be nice. -Susanlesch 05:21, 31 December 2006 (UTC)
I don't know, but striking comment as I see it's still considered "in testing". Gimmetrow 05:30, 31 December 2006 (UTC)
Woops, sorry. At the time I assumed it was an old one. -Susanlesch 05:42, 31 December 2006 (UTC)

Image Caption Problem

I'm a new user and working on the page for my home town Clinton, Iowa. I can't seem to get the caption to display properly or to put a frame around the image of the courthouse. The tag reads [[Image:Clinton_IA_Courthouse.jpg|thumb|The Clinton County Courthouse in downtown Clinton, Iowa.|]] and does not display the caption.

Can someone help out a newbie? I've consulted the Extended Image Syntax page but it does not seem to explain why the caption is not displaying. —The preceding unsigned comment was added by -cman- (talkcontribs) 02:44, 31 December 2006 (UTC).

Fixed. You need to remove the | at the end of the tag, since that causes it to mess up. Tra (Talk) 03:16, 31 December 2006 (UTC)


Thanks very much for your help. Although someone should review the New syntax for images section in the above refrenced Extended Image Syntax section because almost every one of them contains the |]] closing tags.---cman- 06:24, 31 December 2006 (UTC)
The problem is mentioned in that section, in the paragraph starting In particular, if the last option is the void text... but it is a bit easy to miss. The examples given in that section that do end with |]] I think are not intended to have a caption. Tra (Talk) 14:24, 31 December 2006 (UTC)

Little square boxes all over the place.

I've been getting little square boxes all over the place for months now- the kind of boxes that mean that you're missing a font or some such technicala problem and the computer doesn't know how to read them. However, my particular line of editing means I have to be able to read them, so does anyone know what font I need or what I need to do to get them to go away? For instance, here I can't read three of the symbols. Now, if I type in some of the special characters on the bar under the text box I'm typing in now, I get "ḌḍḤḥḶḷḸḹṂṃṆṇṚṛṜṝṢṣṬṭ" which will either look perfectly normal to you because your font works, or it'll look totally wrong to you because your font doesn't work- that's exactly what I'm seeing.

If that made any sense, how does one fix this? Thanatosimii 20:37, 30 December 2006 (UTC)

I can see the characters just fine as a list of consonants with dots under them. I'm guessing it is something to do with your computer, probably missing font support as you suggested. Mike Dillon 20:55, 30 December 2006 (UTC)
See Help:Multilingual support for a list of fonts. And use Firefox, Internet Explorer will usually persistently try to use Arial instead of any Unicode fonts you have installed on your system. If you want to use IE, I suggest installing DejaVu Sans and setting your default font to it under Internet options > Fonts > Latin based. —Ruud 21:36, 30 December 2006 (UTC)
P.S. In IE it sometimes helps to surround the text with {{unicode}}, which forces IE to use an Unicode font if it is installed: ḌḍḤḥḶḷḸḹṂṃṆṇṚṛṜṝṢṣṬṭ —Ruud 21:40, 30 December 2006 (UTC)
where does one get a unicode font? Thanatosimii 22:14, 30 December 2006 (UTC)
You can get DejaVu Sans from their download page. Mike Dillon 22:20, 30 December 2006 (UTC)
DejaVu Sans isn't helping me. Same old boxes. Thanatosimii 22:54, 30 December 2006 (UTC)
What if you add .ns-0 { font-family: "DejaVu Sans"; } to your Special:Mypage/monobook.css and Wikipedia:Bypass your cache? —Ruud 01:04, 31 December 2006 (UTC)
Pardon? I'm not sure how to do that. Do I just type somthing on that page? It doesn't exist yet for me, it says. Thanatosimii 02:50, 31 December 2006 (UTC)
Yep, you've done it right so far. Now press Ctrl + F5 in your browser, which should give a full refresh of the page and make the css code take effect. Tra (Talk) 03:20, 31 December 2006 (UTC)
Afraid I still see boxes... Thanatosimii 03:32, 31 December 2006 (UTC)

loss of editing

I edit an article and make some changes. I then invoke Search to check whether a term in is WP so I can wikify it if appropriate. I then return back to the article I am editing. I find that my edits are lost. The only way to prevent this loss of edits is to do a Preview before I do my Search. My question is: is WP editor supposed to losing edits in this way or is this a software bug that can/should be fixed? Hmains 19:25, 30 December 2006 (UTC)

This is a problem with all HTML forms in general. The best way to deal with it is to use the solutions you have described above. Tra (Talk) 19:37, 30 December 2006 (UTC)
Or open up additional tabs (if your browser supports it) to bring up different pages without losing the form you're on. *Dan T.* 19:49, 30 December 2006 (UTC)
I think this is only a bug in Internet Explorer. If you are using firefox, with caching enabled, you will have no problem with navigating out and not losing your data. Happens to me all the time. GeorgeMoney (talk) 20:30, 30 December 2006 (UTC)

"Undo" feature

Forgive me if this has already been brought up. Just a quick problem, when using the "undo" feature on difference pages. When you click save, if "Prompt me when entering a blank edit summary" is checked in preferences, the page will not be saved and the reminder about entering an edit summary will be shown at the top of the page - it acts as if no edit summary has been entered. — FireFox (talk) 20:01, 27 December 2006

The edit summary prompt checks to see if the edit summary has been changed from the default, not if it's blank. Therefore, it gives an error when you attempt to undo a revision and leave the default summary. You can fix this by putting the following code in your monobook.js:
 //fix edit summary prompt for undo
 addOnloadHook(function () {
   if (document.location.search.indexOf("undo=") != -1
   && document.getElementsByName('wpAutoSummary')[0]) {
     document.getElementsByName('wpAutoSummary')[0].value='';
   }
 })
Tra (Talk) 22:37, 27 December 2006 (UTC)
Huh? I don't have that problem - and I do have "prompt if blank edit summary" enabled. When undo is used and manages to undo the diff, it will generate a semi-automatic summary "Undo version <oldid> by [[Special:Contributions/<user>|<user>]] ([[User talk:<user>|<user>]])" which you can change. If it cannot undo, it will display the current version and no edit summary is provided. Kimchi.sg 01:12, 28 December 2006 (UTC)
I had the same thing happen to me the other day. Odd. EVula // talk // // 19:10, 28 December 2006 (UTC)
I think it could be related to Bug 8340... mind checking if you can replicate it, and if you can, post it there? Titoxd(?!?) 22:55, 29 December 2006 (UTC)
I can't seem to replicate it at the moment - in fact I can't even seem to get the undo feature to work at all at the moment...?? — FireFox (talk) 15:54, 30 December 2006

Disambig link formatting

Currently the templates like {{dablink}} use wiki markup for formatting:

:<div class="dablink">''{{{1}}}''</div>

But shouldn't this really be

<div class="dablink">{{{1}}}</div>

and add this to the site-wide CSS?:

.dablink {
  font-style:italic;
  padding-left:2em;
}

Likewise for other built-in or "meta" messages. — Omegatron 01:16, 20 December 2006 (UTC)

Yes. Actually, it should probably be <p class="dablink">{{{1}}}</p>, although that's arguable. —Simetrical (talk • contribs) 07:23, 21 December 2006 (UTC)
Agreed. How do we get this implemented? Just go ahead and do it? — Omegatron 19:03, 27 December 2006 (UTC)
That tends to be how wikis work, yep. —Simetrical (talk • contribs) 22:29, 28 December 2006 (UTC)
This change cannot be made cleanly. The {{dablink}} template has been subst'd in way too many places. Please revert. Mike Dillon 22:09, 29 December 2006 (UTC)
Please do not change it to <p> tags. That's taking it too far. -- Renesis (talk) 22:46, 29 December 2006 (UTC)
What's wrong with p tags? Is the disambig text not a paragraph? — Omegatron 06:29, 30 December 2006 (UTC)
Is a div or p more appropriate semantically? — Omegatron 07:24, 30 December 2006 (UTC)
I don't know. Either way is arguable. Probably <div>. —Simetrical (talk • contribs) 00:38, 31 December 2006 (UTC)

Citing Images

How do I cite an image taken from Wikipedia? http://en.wikipedia.org/wiki/Image:Charles_de_Gaulle.jpg If suppose I cite as Wikipedia: Charles_de_Gaulle.jpg, typing the same into the search bar wont take me to the image. Help —The preceding unsigned comment was added by 24.215.124.23 (talkcontribs) 17:20, 23 December 2006 (UTC)

You need to type Image:Charles de Gaulle.jpg using "Image" not "Wikipedia". Tocharianne 04:10, 31 December 2006 (UTC)

Template with Parser Functions

I need help from someone well versed in advanced wikimarkup, parser functions and CSS:

http://pt.wikipedia.org/w/index.php?title=Predefini%C3%A7%C3%A3o:Info_animang%C3%A1/Cabe%C3%A7alho/Teste&oldid=4439513

This template is pretty much a version of http://en.wikipedia.org/wiki/Template:Infobox_animanga/Header with variables to control the colors.

If I set the cor_fundo_titulo variable to any CSS color, it works fine, however, if I don't set the variable, it should default to #ccf, but it doesn't. Any clues?

Another problem happens when you paste the first #if: statement in place of the second #ccf string (so the variable can be used to replace its color too). There seems to be a (small) limit in the MediaWiki parser. --ren 06:41, 30 December 2006 (UTC)

The reason is the special | (pipe) syntax used in variable includes. {{{cor_fundo_titulo}}} will literally print out "{{{cor_fundo_titulo}}}" if the variable is not defined. However, if you put a | after the variable name, then whatever is after the | will become the default text to be displayed if the variable is not defined. If you just put {{{cor_fundo_titulo|}}}, then the result is "" and the {{ #if: ...}} statement returns false. Right now, it returns true because there is no default text and the result is "{{{cor_fundo_titulo}}}".
Using this idea of default text, you don't even need to use the "{{{ #if:...}}}" statement. You could replace

{{ #if: {{{cor_fundo_titulo}}} | {{{cor_fundo_titulo}}} | #ccf}}

With this

{{{cor_fundo_titulo|#ccf}}}

Hope that helps! -- Renesis (talk) 07:21, 30 December 2006 (UTC)

Restoring edit history

Hello. Is it possible to restore the edit history from MacHack (chess) to Mac Hack (chess), and Talk:MacHack (chess) to Talk:Mac Hack (chess)? I tried be bold and failed. (Copied the originals to new article, then redirected the originals to the new ones.) Sorry about that. --Susanlesch 06:15, 30 December 2006 (UTC)

Done. For future reference, you can always move the article by using the move tab at the top. Titoxd(?!?) 06:21, 30 December 2006 (UTC)

Someone's screwed up the formatting of the dablinks!!!

<div class="metadata divbox divbox-green" title="CSS updated - try bypassing your cache." >

CSS updated - try bypassing your cache.

They are now officially a mess. Paul Silverman 21:58, 29 December 2006 (UTC)

Any particular examples? User:Zoe|(talk) 21:59, 29 December 2006 (UTC)
See the discussion above and at Template talk:Dablink. This change needs to be reverted. Mike Dillon 22:10, 29 December 2006 (UTC)
The other problem with this is that the "dablink" class is used in many more templates than just {{dablink}}. Mike Dillon 22:13, 29 December 2006 (UTC)

Yeah, look at this one



{{otheruses2}}


there's no formatting in them. Paul Silverman 22:09, 29 December 2006 (UTC)

Relax! We are updating the formatting to CSS. See Template_talk:Dablink#visual_formatting_should_be_done_in_CSS.2C_not_wiki_markupOmegatron 22:40, 29 December 2006 (UTC)

Thanks for commenting here. I was going to follow up. I agree that this is now under control. The only outliers will be cases where the templates were subst'd and not caught in the database dump. Those will need to be caught later and unsubst'd. I think Paul Silverman's problem was actually that he had a cached copy of the CSS. There is also a double-indent issue that is less noticable and is being fixed as we speak. Mike Dillon 22:47, 29 December 2006 (UTC)
There's a question of best implementation, though. Should all the templates be converted to paragraph tags instead of spans and divs, or should the spans and divs be displayed as block elements with CSS? Probably converting all the tags to p class="dablink" is best. 22:50, 29 December 2006 (UTC)
It's fixed. I've put a nasty green box at the top of this section so people purge their browsers' cache, which will solve the issue. (Feel free to take the box out if it looks awful.) Titoxd(?!?) 23:13, 29 December 2006 (UTC)
I think the relevant link is WP:CACHE, not WP:PURGE. Mike Dillon 23:23, 29 December 2006 (UTC)

They are still not formatted!

No. It's still screwed up. THEY ARE NOT FORMATTED. Paul Silverman 23:14, 29 December 2006 (UTC)

Did you do a shift+reload on the page to force the CSS to reload? Mike Dillon 23:22, 29 December 2006 (UTC)

Edit count

Yeah, like how do you find out your own edit count? Max naylor 21:48, 29 December 2006 (UTC)

There are a list of edit counters available here. --MZMcBride 21:49, 29 December 2006 (UTC)
Thank you!!! Max naylor 12:14, 30 December 2006 (UTC)

Proposal : An article title subject to wrapping in a standard browser use because of its length just takes the font size of a section heading. Which is readable as you can see above. See Swpb's question above. What do you think ? -- DLL .. T 08:17, 28 December 2006 (UTC)

Fixing the line-height would be a far more sensible solution than such a hack. —Simetrical (talk • contribs) 22:39, 28 December 2006 (UTC)
I agree with Simetrical - having article titles in different font sizes would be just about as irksome as the original problem. — Swpb talk contribs 22:45, 29 December 2006 (UTC)

A change in the recent changes

I looked at the recent changes and saw that some changes have (+xx) and others have (-xx). What is this, and what do the numbers mean? --AAA! (AAAAAAAAAAAA) 04:21, 23 December 2006 (UTC)

It's the amount of stuff removed or added. -Amarkov blahedits 04:23, 23 December 2006 (UTC)
Amount of characters to be precise :D GeorgeMoney (talk) 04:36, 23 December 2006 (UTC)
The numbers and classes to style them have been there a few days, the actual color change took place in MediaWiki:Monobook.css (more recently). You can copy the styles to your user/monobook.css and change them if you wish. --Splarka (rant) 08:07, 23 December 2006 (UTC)
Amount of characters or amount of bytes? Last time I heard, it was the latter. Titoxd(?!?) 08:15, 23 December 2006 (UTC)
Probably number of bytes, as that coding solution would be quite trivial. --- RockMFR 06:25, 24 December 2006 (UTC)

Bytes. 86.133.54.220 07:58, 24 December 2006 (UTC)

No, I think it's characters. --AAA! (AAAAAAAAAAAA) 12:22, 24 December 2006 (UTC)
Every character is a byte long, so in this case I guess it'll mean the same thing. From a programmer's point of view, a counter of bytes will do the trick. — 0612 (TALK); Posted: 13:14, 24 December 2006 (UTC)
It's bytes. See Article.php, lines 1280 and 1287, $oldsize and $newsize set using strlen. Most characters on English wikis are one byte long, but there will be a noticeable difference on almost all other wikis, and it will be very noticeable on non-Latin wikis. —Simetrical (talk • contribs) 14:50, 24 December 2006 (UTC)
Hmm... I do believe you, but it says here it means characters. --AAA! (AAAAAAAAAAAA) 00:58, 26 December 2006 (UTC)
I agree with Simetrical, it's bytes. I've updated the page accordingly. Tra (Talk) 01:28, 26 December 2006 (UTC)

It's pointless. Here's the closest thing we have to documentation, though.

I would like the numbers to be removed and this functionality provide a a short summary of the edit instead. Why make things mysterious? — Omegatron 06:44, 30 December 2006 (UTC)

Could a developer update Special:Shortpages ? The current version was cached on 4 December and is totally useless now. I know this is supposed to be quite ressource intensive, but it would be nice to have a regular and, most importantly, predictable update frequency — the more often the better (although less than 2 days apart would be too much anyway since it takes some time to process all entries), but I don't know what would be reasonable from the technical point of view. Schutz 12:08, 20 December 2006 (UTC)

Try these links: Wikipedia talk:Special:Shortpages & User:Zorglbot/Shortpages.--NMajdantalk 22:04, 20 December 2006 (UTC)
Hum... I have left a message on Wikipedia talk:Special:Shortpages a few weeks ago already, and User:Zorglbot/Shortpages is created and updated (from Special:Shortpages) by my own bot... Thanks anyway, Schutz 00:07, 21 December 2006 (UTC)

It seems that all the Special: pages that cache lists of pages haven't been updated since 4 December, or earlier in some cases. Until recently, they have all been updating regularly every 2-3 days. Does anyone know what has happened? --Russ (talk) 11:46, 21 December 2006 (UTC)

There have been multiple complaints about this on Bugzilla, so I can assume the devs are looking into it. --ais523 16:37, 22 December 2006 (UTC)
Special:Uncategorizedcategories has the same problem, last updated December 4th. -- ProveIt (talk) 13:31, 23 December 2006 (UTC)
Do you have bug IDs so that I can follow the discussion ? I haven't found them after a quick search. Schutz 15:58, 23 December 2006 (UTC)
It appears to be this one. -- JLaTondre 16:22, 24 December 2006 (UTC)
Thanks. Schutz 20:59, 25 December 2006 (UTC)
  • All of these queries, across all Wikimedia wikis, have been stopped, mostly due to lack of processing power on the database servers. (They can only handle so much load, you know...) The best chance of them getting up soon is if you help bring in some more badly-needed servers... Titoxd(?!?) 08:51, 27 December 2006 (UTC)
Ok, this is good to know — I have deactivated my bot which parses Special:Shortpages — if the cached page does not change, there is not much point harassing the database server by requesting 1000 pages and writing 1 every day... It would be nice it these pages could be updated, even every 2 weeks or more, but I understand that there is only so much processing power. Schutz 10:40, 27 December 2006 (UTC)
Special:Shortpages has been updated, yeah ! Schutz 16:24, 29 December 2006 (UTC)

Quicktime

When I click on a link for more information I get a page in quicktime which I cannot read. —The preceding unsigned comment was added by 138.130.45.100 (talk) 12:47, 29 December 2006 (UTC).

Where? 86.133.54.220 14:53, 29 December 2006 (UTC)

evansdale iowa action req'd

This page has a "profanity" onNankai 06:15, 29 December 2006 (UTC) it, my public library access is denied as a result. I can't even see what the profanity is! Can anyone fix? Is the profanity required? Nankai 06:15, 29 December 2006 (UTC)

The page hasn't been changed recently, and there is nothing on it that looks like profanity to me. I'm guessing your library system is wound a little too tight. Dragons flight 08:12, 29 December 2006 (UTC)

Links

Is it just me, or are all categories showing up as red links... Even if you check the main page, under the "In Other Languages" header, all the links are the language code (i.e. en: es: it: etc.).

Thanks.

(As I'm saving this, it says: the database is locked for maintenance, so, maybe this problem will be fixed as soon as I click "Save Page")

Tyson Moore es 01:53, 29 December 2006 (UTC)

Yep. That fixed it alright....
Tyson Moore es 01:54, 29 December 2006 (UTC)

The file we use to store interwiki information became corrupted for some reason around 01:40 - 01:55 this morning when a new wiki was added to the cluster; this was corrected soon after, but caused some problems for a brief period. 86.133.54.220 14:55, 29 December 2006 (UTC)

Image help

Need some technical assistance with Image:Ontario-timmins.PNG. I added a city infobox to the Timmins, Ontario article recently, but the map image won't show up in the template. Apparently the 250px thumbnail (the one that the infobox template actually looks for) is broken in this particular case (see [3]), although other image sizes work correctly. Is there something the technical team can do to resolve this, or should I just try deleting the image and reuploading it? Bearcat 01:41, 29 December 2006 (UTC)

Fixed. Thanks to those of you who helped. Bearcat 03:13, 29 December 2006 (UTC)

Version of librsvg ?

I have some problems with two SVG images that I uploaded on commons (namely Image:Dotplot-example.svg and Image:Michelsonmorley-boxplot.svg). If I understood well, librsvg is used to rasterise these images. Is there a way to know which version is used ? I have no problem using librsvg 2.14, so I wanted to see which version is used by Wikipedia to try to pinpoint where the problem is (upstream or in Wikipedia). Thanks, Schutz 00:34, 29 December 2006 (UTC)

We're using a slightly patched 2.14.0 at the moment. --brion 01:50, 29 December 2006 (UTC)

'Enhanced recent changes': never see expansion arrows?

Hi there. I've been messing around for some time, trying to get the nice 'enhanced recent changes' I see on a very few meta-watch pages that use that template, for use on my own watchlist.

I've found that, in fact, I'm using it (preferences have said that I have for a while), but when I look at source for my watchlist, every entry looks like this:

img src="/skins-1.5/common/images/Arr_.png" (blank arrow) instead of
img src="/skins-1.5/common/images/Arr_r.png

That is to say, no matter what watchlist entry I'm looking at, it's giving me a blank arrow, and won't let me look at multiple recent changes with a mouse click. I've made sure I'm looking at multiple days worth of entries, so I think I should be getting the expansion arrows. Is something broken with my 'Enhanced recent changes' template, or perhaps all enhanced recent changes user templates? Could I be missing something obvious? Skybunny 17:30, 28 December 2006 (UTC)

Try going to my preferences > Watchlist and tick 'Expand watchlist to show all applicable changes'. Tra (Talk) 17:51, 28 December 2006 (UTC)
That was it exactly. I admit, I didn't really see the connection between 'Enhanced recent changes (JavaScript)' on the 'Recent Changes' tab, and 'Expand watchlist to show all applicable changes' on the 'Watchlist' tab. I didn't realize they were related. (Should they be on the same tab or somehow rephrased to make their meaning clearer? I don't know.) Anyway, thanks much. Skybunny 18:02, 28 December 2006 (UTC)

Boldness of headings

Level 3 headings are being displayed bolder than level 2 headings (on my computer, etc). I use the default skin. Is this a setting or do developers need to be invovled to get it changed? RJFJR 17:06, 28 December 2006 (UTC)

It can be changed by an admin in MediaWiki:Monobook.css. I don't think there's a problem with this since level 2 headings are distinguished by their larger font size and underlining. Tra (Talk) 17:17, 28 December 2006 (UTC)
They don't look bigger to me. They look the same size but L3 isn't underlined but is bolded. RJFJR 18:12, 28 December 2006 (UTC)
Level 2 headings are slightly bigger. Here's the part of the stylesheet http://en.wikipedia.org/skins-1.5/monobook/main.css?37 where the headings are formatted:
h1, h2, h3, h4, h5, h6 {
	color: black;
	background: none;
	font-weight: normal;
	margin: 0;
	padding-top: .5em;
	padding-bottom: .17em;
	border-bottom: 1px solid #aaa;
}
h1 { font-size: 188%; }
h1 .editsection { font-size: 53%; }
h2 { font-size: 150%; }
h2 .editsection { font-size: 67%; }
h3, h4, h5, h6 {
	border-bottom: none;
	font-weight: bold;
}
h3 { font-size: 132%; }
h3 .editsection { font-size: 76%; font-weight: normal; }
h4 { font-size: 116%; }
h4 .editsection { font-size: 86%; font-weight: normal; }
h5 { font-size: 100%; }
h5 .editsection { font-weight: normal; }
h6 { font-size: 80%;  }
h6 .editsection { font-size: 125%; font-weight: normal; }

.editsection {
	float: right;
	margin-left: 5px;
}
As you can see, level 2 headings are sized at 150% but level 3 headings are sixed at 132%. Tra (Talk) 18:29, 28 December 2006 (UTC)
You can change this for yourself if it really bothers you. Add the following to User:RJFJR/monobook.css:
h1 { font-weight: bold }
h1 .editsection { font-weight: normal }

h2 { font-weight: bold }
h2 .editsection { font-weight: normal }
That will make first and second level headings bold, just like h3 and lower. The second statement in each pair prevents "[edit]" from being bolded too. Mike Dillon 18:34, 28 December 2006 (UTC)
Thank you! RJFJR 19:20, 28 December 2006 (UTC)

Numbers next to username.

What exactly are the green +ve and red -ve numbers next to usernames on pages like my watchlist and edit histories?BigHairRef | Talk 02:04, 28 December 2006 (UTC)

They represent the change in character count between the most recent and the next most recent edit. Wahkeenah 02:16, 28 December 2006 (UTC)
Well, it's actually the amount of bytes, not characters. (They are almost the same thing, but bytes are slightly smaller than characters, I believe.) –The Great Llamasign here 19:49, 28 December 2006 (UTC)
The number of bytes will always be greater than or equal to the number of characters. This is because the minimum number of bytes to represent a character is one byte, but some characters need more than one byte in the database. No characters require less than one byte for storage. Mike Dillon 21:48, 28 December 2006 (UTC)
See Wikipedia:Watchlist#What do the colored numbers mean?The Great Llamasign here 02:06, 28 December 2006 (UTC)

Error with image

http://en.wikipedia.org/wiki/Image:Lou_p.gif

The following error message appears:

(56KB, MIME type: image/gif)

[[[[[[Image:Failed to parse (unknown error): --(four ~ appear here, but don't sign a name) 12:46, 27 December 2006 (UTC)#REDIRECT [[Media:'Bold text == Summary == Pasteur and his device for disproving spontaneous generation theory. Which he did.]] ]]]]]]

Please could someone fix this minor fault. Thanks Micro.pw 12:46, 27 December 2006 (UTC)

The image description page was vandalised. I've fixed it now. Tra (Talk) 13:15, 27 December 2006 (UTC)

Florida router down?

According to the Internet Traffic Report, a major router at Ft. Lauderdale is down. My connection to Wikipedia has been horribly slow for prolonged periods of time, and has produced an innumerable amount of session losses; at times, the site is completely unreachable, even though ICMP pings return fine. The odd thing is that the rest of the Internet looks OK to me as I browse. So, could it be related to this? How can I switch to the Kennisnet proxies, at least for page reads, to see if this is the issue? Titoxd(?!?) 08:48, 27 December 2006 (UTC)

You can set proxies in your web browser or computer's configuration (details may depend). European proxies are at rr.knams.wikimedia.org. Note that unless you have some clever browser plugin, switching them in will make everything _else_ on the web fail to work, so use with caution. ;) --brion 09:51, 27 December 2006 (UTC)

I think there is a global problem with internet traffic ever since the Taiwan earthquake.[4] See: http://www.internettrafficreport.com/7day.htm for a chart showing how traffic is down and delays are up. Hu 02:59, 28 December 2006 (UTC)

Those cute tinted little text boxes - the kind with dashed lines to frame them

I need to know how to make these. My edits in a very long article had been up only 12-15 hours when they all disappeared into inline something or other. I forget what she said they were called. It is when there is < ! --- write stuff here --- >. I am supposed to find these entire paragraphs and fix them with citations and rewriting or they will be deleted. I don't know if this would upset any bots or trouble-shooters, but I am HOPING to make my stuff re-appear so I can find and edit it ... and maybe even one of the users can give me some ideas. Can anyone here help me?

It is increditbly difficult to find them --Kiwi 00:10, 27 December 2006 (UTC)

Your text is currently at the beginning of the "Current Pharmacology Treatments" section of Borderline personality disorder. You can see it by using this edit link and removing the "<!--" at the beginning and the "-->" at the end of your text. Please use "Show preview" to view the text until the citation concerns are addressed. Mike Dillon 00:33, 27 December 2006 (UTC)
Thank you, Mike. Actually, there are 4 paragraphs they were done to and the rules say not to change edits for several days until people can discuss the editing changes. I don't want to get on the wrong side of anyone. I had proposed this (on the talk page in response to the announcement of what had been done "as an experiment", but only my stuff is being experimented with. See what I mean?
And what do you mean using "how preview? Does this mean I have to put back the hidden stuff after I do any changes until it is perfect? I don't know enough to be perfect. And there are lots of missing citations that have been up for a long time, and mine half a day. I guess I just don't understand why this happened. I don't think I like this experiment. It makes things hard. --Kiwi 00:43, 27 December 2006 (UTC)
I said "Show preview", referring to the "Show preview" button on every edit page.
As for the "experiment" happening with Borderline personality disorder, this is not a technical issue. Please discuss it with the involved editors at Talk:Borderline personality disorder. Mike Dillon 00:48, 27 December 2006 (UTC)
So I can't post anything until it's perfect. Just look at it in preview. Okay.
And I can't undo the experiment until that person comes back to Wiki - which might not be for several days. And if that person says, "No", then I'm STUCK???? Is this my only way out of this? This is a bummer.
If I could let people see what I wrote, I could get feedback on it from people who don't necessarily check out the talk page. --Kiwi
Forgot to mention that I posted a detailed response to the person who decided to create this experiment early this morning, but nothing so far. Not even other people. --Kiwi 01:02, 27 December 2006 (UTC)
As I said, none of this is a technical issue. I didn't mean to imply that you must use "Show preview", it was just a suggestion to avoid conflicts while working to source your additions. I would suggest that you go ahead with whatever edits you feel are appropriate; since there have already been objections to your additions, it would be prudent to continue to explain your changes on the talk page as you have already done. I don't see any particular urgency to your changes and you should bear in mind that this is a holiday period and many editors are not active. I'd suggest being patient. Mike Dillon 01:11, 27 December 2006 (UTC)

Status of secure.wikimedia.org?

I've noticed that a lot of people are being pointed to the secure.wikimedia.org address to bypass proxy problems. I myself use it for security paranoia reasons (I edit over a number of unsecured wireless networks) and have noticed that not only is the connection slower (as noted in this page's header), but that it occasionally returns sporadic 404s that go away on refresh and has some problems with URL rewriting (especially article titles with question marks; try Life on Mars and Life on Mars?). Is secure.wikimedia.org just a hack to work around proxies, and only that, or is it an "official" HTTPS server that is intended to be just like en.wikipedia.org (in which case I'd lobby for https://en.wikipedia.org/ as the address)? æ²  2006‑12‑26t22:47z

It is official, since it's under the wikimedia.org domain. It is also, as you describe it, a hack to work around proxies. However, if you are using it for security reasons, you may wish to be aware that it's not designed to be completely secure, e.g. it uses null encryption for performance reasons(not true, see below). Your idea for givng it the url of https://en.wikipedia.org/ is a good one and, if it can be technically implemented, it would be useful since it would be easier to remember because of its similarity to the non-secure URL. Tra (Talk) 23:41, 26 December 2006 (UTC)
Is the use of the null cipher documented anywhere? My browser says that it's AES 256 as far as I can tell and the stream I captured with Wireshark seems to be something other than plain text. Mike Dillon 00:25, 27 December 2006 (UTC)
I remember reading it somewhere but I can't remember where. Maybe they do use encryption? Tra (Talk) 01:19, 27 December 2006 (UTC)
I mainly asked because someone was asking about HTTPS encryption on Meta in regards to the MediaWiki API. I'd be interested to know if it is actually unencrypted. Mike Dillon 01:23, 27 December 2006 (UTC)
It is not documented, probably because it is not true. :) --brion 04:53, 27 December 2006 (UTC)
IIRC (this is all from memory), the problem with it being at https://en.wikipedia.org/ is that the certificate used is for secure.wikimedia.org, which would cause a security warning from the browser due to the mismatch. The server cannot present a different certificate because the requested hostname is sent after the server must send its certificate. Newer browsers are able to send the requested hostname in the initial handshake (before the server sends its own certificate), but requiring that would restrict which browsers can be used. There's also the problem that secure.wikimedia.org is a completely different machine than the rest of the Apache servers (which is why it's much slower...), and https://en.wikipedia.org/ would point to the later, instead of to the former. --cesarb 02:18, 27 December 2006 (UTC)
The funny thing about that is that the certificate will generate a warning on most browsers anyways since CAcert.org is not a recognized Root Certificate by most browsers (as far as I know). Mike Dillon 02:33, 27 December 2006 (UTC)

The present secure.wikimedia.org is experimental; it may be altered or discontinued at any time (though it is more likely to be altered than discontinued).

Currently:

  • it's relatively slow, served out of a single server
  • it uses a funny URL structure which kind of sucks
  • it uses a certificate signed by a non-default signing authority (CACert), which will toss up warnings on most browsers the first time you visit it
  • there may be some inconsistencies due to running on a different version of Apache from other servers

Eventually it will probably be reorganized, possibly in a nicer fashion using the main hostnames if we can work out the proxy and virtual hosting setup. When we actually think it's reasonable to use for the general public we'll most likely also get a "real" certificate from a CA that doesn't toss up warnings on everybody's browsers. --brion 04:51, 27 December 2006 (UTC)

Hughesnet users unable to log in?

An editor who uses Hughesnet has claimed that WP "sign in software" does not work with this ISP. [5] This is only really an issue because Bob (that's how he usually signs himself) operating from this dynamic IP has made some very controversial edits to pages concerning Marc Lepine. (a few diffs: [6] [7].) I'm inclined to think he's telling the truth as the article was sprotected for a few weeks and he did not edit it. However, its difficult to warn, converse, or block this user because his IP changes quite rapidly. This issue must have come up before with Hughenest -- is there a way for him to log in? He's not a vandal, just a POV pusher and it would be easier to deal with him if he could log in. I expect he would create an account if the option were presented -- he gets rather annoyed at being referred to as an "anon". It doesn't really begin to solve the problems at that article, but it's a start. Anyone dealt with this before? Dina 22:37, 26 December 2006 (UTC)

Try asking him to use the secure server at https://secure.wikimedia.org/wikipedia/en/wiki/Main_Page. Tra (Talk) 22:44, 26 December 2006 (UTC)
Thanks! Dina 23:05, 26 December 2006 (UTC)
I'm attempting to contact HughesNet to see if we can resolve this problem once and for all... they don't make it easy to find an appropriate contact point, though, so I don't know if we'll get anywhere. :P --brion 05:21, 27 December 2006 (UTC)

Title formatting

Sometimes unavoidably long article titles will wrap to a second line. For some reason, the vertical spacing between these lines is very small, which looks terrible. I was wondering if any other wikipedians are bothered by this, if there is a reason for it, and if not, how it can be fixed. —Swpb talk contribs 20:25, 26 December 2006 (UTC)

Did you try "Category:Articles whose titles are too long" ? some titles are just nicely truncated. When the Pawn is my fav'. -- DLL .. T 21:51, 26 December 2006 (UTC)
The issue is not one of the title being over the 256 character limit. Case in point, the article Experimental Assembly of Structures in EVA and Assembly Concept for Construction of Erectable Space Structure, whose title must be written out per Wikipedia:Naming conventions (acronyms)Swpb talk contribs 22:14, 26 December 2006 (UTC)
Try placing .firstHeading { line-height: 1.25em; } in your user CSS. æ²  2006‑12‑26t22:50z
This is due to a bug in the MediaWiki monobook skin's css (specifically in main.css). A line-height should not be defined in terms of em for any element that has child elements. This had been briefly fixed but had to be reverted because it very subtly changed the overall monobook stylings too much. There was an old thread here about this too. --Splarka (rant) 08:24, 27 December 2006 (UTC)
Hello Swpb. Now you've written up one case, (for which Ease Access should be enough) and the current guideline (which is not compelling), I'm able to tell that the visual impression is indeed not too good.
The solution must not be found in any user css. What would just adding a newline in the title do : can it be done ? -- DLL .. T 18:44, 27 December 2006 (UTC)
In fact you do not know where wrapping occurs, it depends on too many things, e.g. showing history or bookmarks in a left panel. Maybe a little educated guess : if the title is long enough to get wrapped, maybe more than 40 chars ... reduce the font size (and not the line height). Would that do ? -- DLL .. T 18:49, 27 December 2006 (UTC)
I'd like to see something done about this in the main css, rather than just my personal css. I read over the old thread Splarka linked to, but the coding jargon was a bit over my head, and the thread seemed inconclusive. It seems (to a code illiterate person like myself) like there should be a way to fix this without disturbing other style elements. Does anyone know how I would go about proposing this to the powers that be? —Swpb talk contribs 20:15, 27 December 2006 (UTC)
Someone just has to get around to figuring out the right incantations so stuff doesn't get screwed up. It's been on my to-do list for a long time . . . I'll get around to it sometime. —Simetrical (talk • contribs) 22:34, 28 December 2006 (UTC)

Infobox appears to be automatically generating a link to a dab page

As part of my disambiguation work (I've adopted English, among other pages, and am trying to keep it clear of incoming links from mainspace), I'm having trouble with this article. As far as I can tell, the non-linked text "English" in the {{Infobox Former Country}} is somehow getting turned into a link by the infobox template, resulting in a link to the dab page. I'd like to disambiguate the link, but I don't know how to do so when the link is being automatically generated by a template. Any help? --Tkynerd 14:00, 26 December 2006 (UTC)

Looking at the template code, it simply turns the text into a link if an article by that name exists, nothing more. If you put a link there instead of the text, it will confuse the code enough that it'll leave you alone (for instance, changing it from English to [[English language|English]] will do as expected). I'd recommend, however, that you also track whoever makes these templates and ask for suggestions; they might be able to add even more magic to the code to make it automagically do whatever is needed. --cesarb 14:59, 26 December 2006 (UTC)
Thank you! Now I can finally get this article off my little list of articles that can't or shouldn't be fixed. (For English there are two that shouldn't be fixed...and then there's this one.) I'll try to get in touch with whoever is responsible for this template to see if there is anything that can be done about the problem. --Tkynerd 15:15, 26 December 2006 (UTC)

Image/SVG trouble?

I've recently uploaded Image:VISA_Logo.svg, but it doesn't show. It won't show with a smaller size either, like in this template:

A vector version of this image is also available, and should be used in place of this raster image whenever possible.
For more information about vector graphics, see the articles on vector and raster graphics. There is also information about MediaWiki's support of SVG images.

Village pump (technical)/Archive AF

What's going on? Watching the actual SVG works both in Firefox, Opera and Konqueror.

DarkPhoenix 11:38, 26 December 2006 (UTC)

http://upload.wikimedia.org/wikipedia/en/thumb/e/ed/VISA_Logo.svg/800px-VISA_Logo.svg.png was not generated. -- ReyBrujo 12:00, 26 December 2006 (UTC)
Yeah, well, I saw that myself. Question is: Why? DarkPhoenix 12:09, 26 December 2006 (UTC)
It works now. It wasn't working because the SVG code was so messy and there was an external link. I saved it as plain SVG and did some code cleanup. --Uusitunnus 14:27, 26 December 2006 (UTC)

Renaming A File

Many of us agree that the file name of the image for the Kent State shootings is not NPOV. Is there a way we can rename it?

--Knulclunk 05:02, 26 December 2006 (UTC)
I have to interject here: this comment is misleading. There has been no discussion of this on the article talk page whatever. If "many of us agree" about it, they are invisible, because the only comments about this (other than one I just made) are on the photograph's talkpage, and they consist of one user in May 2005, one IP address in April 2006, and the above user. That is hardly "many of us". In any case the person who uploaded the picture is the one who chose the name, which is in fact a phrase that is used on the article page itself to describe the event, and it's no more POV than to call JFK's death an assassination rather than a death. Tvoz 07:57, 26 December 2006 (UTC)
A quick check shows that the term "Kent State massacre" is a common characterization of the event in sources that aren't pushing a detectable POV (though a strong argument can be made that the word "massacre" is formally incorrect in this case). Since the term is in common usage it should be acceptable here. Raymond Arritt 08:09, 26 December 2006 (UTC)
The Boston Massacre only had five dead. --Chris Griswold () 10:33, 26 December 2006 (UTC)


I thought this was a technical forum. But, if you need to filter the value of my question:

1.The act or an instance of killing a large number of humans indiscriminately and cruelly.
2.The slaughter of a large number of animals.
  • massacre from Wordnet:
the savage and excessive killing of many people
  • massacre from Wikipedia:
The word massacre has a number of meanings, but most commonly refers to individual events of deliberate and direct mass killing, especially of noncombatant civilians or other innocents without any reasonable means of defense, that would often qualify as war crimes or atrocities...Additionally, the word massacre is often used for political or propaganda purposes, and the choice of whether to label an event a massacre may become a sensitive one; see, for example, the Kent State massacre

A truly neutral(NPOV) title for this picture would be a string of numbers. I don't have an agenda here, just trying to make Wikipedia a better place.
Can anyone help with the technical aspects of the question and we can move the rest of this to a talk page?
--Knulclunk 21:56, 26 December 2006 (UTC)

From a technical point of view, to rename an image, you will need to:
  • Save it to disk
  • Upload it under the new name
  • Copy and paste the image description to the new image page
  • Update all page references to the image
  • Tag the old image for speedy deletion under WP:CSD#I1.
Tra (Talk) 22:17, 26 December 2006 (UTC)

Thanks! -- Knulclunk 21:22, 27 December 2006 (UTC)

Performance Issues

2 days ago, I started to experience severe performance issues with the site. Pages seem to take forever to display, even short ones. Long articles and pages like the Helpdesk, may get cut off with only part of the page being displayed. It even seems to take a long time to display "my contributions". Also, when I try to "save page" I frequently get an error that there is a loss of session data. The problem does not seem related to my browser, since I am having the same problem with both IE 7 and also Firefox (which I have just now installed). I also don't think that it is my monobook or popups, since blanking monobook.js doesn't seem to help (yes I've cleared my cache). I don’t think I have a network issue, since a test (at cnet) showed my connection ~900kbps. I have seen some other postings from people having problems, but don't see any solutions anywhere. Can anyone provide any insight here? --After Midnight 0001 19:11, 25 December 2006 (UTC)

Do you have some firewall or antivirus that might be slowing wikipedia down? Do other sites load slowly? It could just be that it is holidays and everyone is on the internet, causing traffic for your provider. GeorgeMoney (talk) 20:18, 25 December 2006 (UTC)
Other sites seem fine, except for the other language wikipedias (like Simple English and German) also seem slow. I can't imagine that it would be a firewall or AV issue, since I haven't changed anything recently, but of course, these things tend to auto-update these days. I even tried a rollback to a system restore point from before the problem started, to no avail. Do you have any specific things or settings that you would recommend that I check? --After Midnight 0001 20:47, 25 December 2006 (UTC)
Another editor experienced similar symptoms recently - most pages taking 10+ minutes to load. Solution was to change the firewall. Gimmetrow 20:52, 25 December 2006 (UTC)
I had to go onto the page history to see what you were referring to.... OK, I've gone into my McAfee and reset the firewall back to defaults. That appears to let me now load some very big pages, like the Help Desk very quickly, but opening up some things, like editing or my contributions still takes a while. And now the Help Desk page is giving me intermittent response again. I still think that there may be something else going on here. --After Midnight 0001 00:29, 26 December 2006 (UTC)
  • No firewall (disabled it a while ago, to try to check if it was the culprit), and I have had the exact same problems on my end as well. I have had severe lag for about two or three days now, I cannot edit at all, as it takes me about ten minutes to try to get an edit in. I also end up losing my editing session a lot, and when combined with Bug 8340, it has proved to be a very miserable editing experience for me as of late. I've checked everything, the rest of the Internet works well, but I have severe problems reaching all Wikimedia sites. It's something either on Wikimedia's end, or upstream from there. Titoxd(?!?) 08:57, 27 December 2006 (UTC)

Replacing names of redirecting templates with target template names

I think it would be useful and would allow for more consistency to have the parser automatically replace template names (in the wiki source) that redirect to other templates with the names of target templates. This would reduce the mess when renaming templates. For example: if an editors puts {{a}} in the wiki source and the 'template:a' is a redirect to 'template:b', then the parser should automatically replace {{a}} with {{b}} in the wiki source. This could be done when the template is renamed or when the page using {{a}} is saved.

Any thoughts? --Eleassar my talk 13:18, 25 December 2006 (UTC)

This is a bad idea. It will result in tons of invisible changes every time a template is changed to redirect anywhere. So if Template:A redirects to Template:B, and then someone changes it so Template:B redirects to Template:A instead, diffs will show people changing {{b}} to {{a}} willy-nilly all over the place. Trying to do it when the redirect itself is established is even worse. —Simetrical (talk • contribs) 22:40, 25 December 2006 (UTC)

Well, you're right but I just hate seeing people and bots waste their time on doing this. Do we have any alternative here? --Eleassar my talk 11:02, 26 December 2006 (UTC)

149px works but 150px doesn't

Why does 149px work but 150px doesn't? 149px:   150px:  

Hopefully everyone else is seeing a broken image for 150px... My guess is a problem occurred when the 150px version of the Commons image was first generated. But how can a regeneration be forced now? —Wknight94 (talk) 19:54, 24 December 2006 (UTC)

I attempted to use "action=purge" as suggested at the top of this page, but it did not do anything. Here's what's actually downloaded for the 150px image:
<html><head>
        <title>Bad title</title>
        <body>
<h1>Bad title</h1>
<p>The requested page title was invalid, empty, or an incorrectly linked inter-language or inter-wiki title.
It may contain one more characters which cannot be used in titles.</p>
</body></html>
The top of this page doesn't say what to do if "action=purge" doesn't work... Mike Dillon 20:40, 24 December 2006 (UTC)
I can't see the 150px either! This happened to me a while ago but it was an SVG file that didn't show when it was at 300px or more, I coudn't figure it out back then and I was just content with 299px! Orionist 20:46, 24 December 2006 (UTC)
Done. You need to copy the url and try to access the page, then append a 1 after the name, so that the thumbnail is regenerated, and then remove it and purge the cache. -- ReyBrujo 20:47, 24 December 2006 (UTC)
Wow. Is that worth reporting in Bugzilla or whatever? Or has it been addressed before? (BTW, thanks for figuring that out!  ) —Wknight94 (talk) 23:04, 24 December 2006 (UTC)
It may have already been reported. Someone had the same problem with an infobox album picture, and asking around, I got the answer from here. It is a problem where the thumbnail is incorrectly generated and, since cached, it won't be regenerated unless you manually force it. It happens often, so it is a handy trick to learn. -- ReyBrujo 04:06, 25 December 2006 (UTC)
I've hopefully fixed this, but I'm not 100% sure. The cache system for thumbnails was hacked together in a hurry a couple months ago when we had serious load problems, and it's... a little ugly. :)
There may still be some bad file entries... ?action=purge _should_ fix them but I'm not convinced it all works.
Note that commons images are run through the extra caching layer, but ones local to en.wikipedia aren't, if I understand the setup correctly. --brion 06:10, 25 December 2006 (UTC)

Problems w/ Popups and Recent Changes

I have the following problems:

  • I'm getting this notice when I use Voice of All's javascript and no one has reverted a user:

Page rollback error: Last editor is <>, not <>. The user may have already been reverted. I waited about 5 minutes, thinking the server may need to catch up or something, but no one showed up in the history as reverting a user

  • The live lists of all recent changes, filter recent changes, and recent IP edits aren't working for me. It just keeps adding another yellow bar every 30 seconds saying it's updating, but no pages show
  • I go to the regular recent changes (the one in the navigation sidebar) to do RC patrolling, I don't even see any recent changes either.

Anyone know why? TeckWizTalkContribs@ 17:51, 24 December 2006 (UTC)

Disabling byte count display

Hi. Please take a look at my monobook.js. Why is my watchlist still showing the byte counts, albeit without colors or bold? Thanks a lot. Xiner (talk, email) 16:45, 24 December 2006 (UTC)

Because it's supposed to be in your monobook.css and what I've done should have fixed it. [8] --Deskana (talk) 16:49, 24 December 2006 (UTC)
Hmm, some changes are still showing numbers on my watchlist. I can't find a pattern to what's happening though. Could you take a look again? Thanks a lot. Xiner (talk, email) 18:46, 24 December 2006 (UTC)
Have you purged your cache (in example, CTRL+F5 in most browsers)? -- ReyBrujo 18:50, 24 December 2006 (UTC)
Yes, I've tried every which way, including going into the browser and clearing my cache. No luck though. Xiner (talk, email) 23:42, 24 December 2006 (UTC)
Ok, it looks like a cache problem. I opened the page in a new Firefox profile as well as IE, and they both display without the numbers. I don't understand why my cache isn't clearing though. Xiner (talk, email) 03:25, 25 December 2006 (UTC)
I didn't try setting the cache to 0 KB, which solved the problem. Thanks. Xiner (talk, email) 03:32, 25 December 2006 (UTC)

Massive main page vandalism

I don't know if this is place to report it, but a massive vandalism of the main page has occured, and I don't know how to revert it. Some neanderthal managed to post a pretty disgusting picture and have that dominate the main page. I tried to revert it, but it keeps showing up on various past versions... — Preceding unsigned comment added by 65.94.175.68 (talkcontribs) 15:39, 24 December 2006

See the discussion Wikipedia:Administrators'_noticeboard/Incidents#Vandalism on Main Page, and please report vandalism like this to the Administrators' incidents noticeboard in future. Thanks. Carcharoth 16:41, 24 December 2006 (UTC)

Image names with punctuation bug

I have tried reporting this bug but it is still happening. If you use punctuation in an image name (ampersand, plus, etc.), the smaller resolution images generated from it will show up then be deleted, never to be seen again. Even a dash seems to cause problems. Please somebody do whatever it takes to get this bug addressed, reporting it does not seem to solve anything. Is anyone with some influence/whatever in these matters willing to make getting this fixed a personal project? Please? Fourdee 14:48, 24 December 2006 (UTC)

What is the bug number for the report at bugzilla? --cesarb 16:30, 24 December 2006 (UTC)
The bug number at bugzilla is 8367. I also had a previous bug report at bugzilla (7762) which did not mention what I now know which is that the problem is due to punctuation in the file name. Fourdee 17:35, 24 December 2006 (UTC)
Oddly, I managed to regenerate the image page thumbnail (the 800px one) for the one reported in bug 8367, and both sample thumbnails in the image talk page for the one reported in bug 7762 are loading correctly. I believe you probably are hitting the same old broken thumbnail problem, and it happening on the images with strange characters in the filename is just a coincidence. However, I found some bizarre cache behaviour when checking; I will add it to the bug report. --cesarb 23:17, 24 December 2006 (UTC)
Also, using the insane workaround suggested below (adding 1 to the end of the URL — WTF?), I regenerated the 250px thumbnail. This only gets weirder. --cesarb 00:05, 25 December 2006 (UTC)

Template transclusion preview proposal

The recent rash of template vandalism takes major advantage of the fact that there is no way to quickly see the rendering of <includeonly> text on a template, unlike <noinclude> or ordinary text. Would the devs please consider an additional view for templates that shows their transcluded appearance? Maybe even two different views depending on whether the template is subst'ed or not.

I've seen "make inline image use within <includeonly> text show up in an image's file links" being proposed as a solution, but I believe a "includeonly preview" mode would be more useful if it is technically feasible, as it can also be useful for ordinary template testing. Mentioning here to see what others think. ;) Kimchi.sg 10:52, 24 December 2006 (UTC)

That sounds like a good idea that I would endorse. I have no idea if the MediaWiki software can currently support it, though. Stevie is the man! TalkWork 05:30, 28 December 2006 (UTC)

SVG renders sigma, but not mu.

I recreated Image:Normal distribution pdf.png as an SVG; see Normal distribution pdf.svg. The μ's aren't displaying, even though the SVG source shows, for instance, <text>μ = 0, σ² = 0.2</text>. It also renders fine in inkscape/inkview. Should I file this as a bug on bugzilla? grendel|khan 07:04, 24 December 2006 (UTC)

Addendum: There's a similar bug 3769, which may be related, but then again, may not be. Perhaps this is an issue with the fonts used in Mediawiki's rasterizer? grendel|khan 07:11, 24 December 2006 (UTC)

in order to avoid this problems, don't use Greek letters at all, but introduce them as pictures. They are already rendered properly here on Commons. Copy, paste and resize the ones you need. For more information and guidelines, see Wikipedia:How to create graphs for Wikipedia articles Alessio Damato 13:18, 24 December 2006 (UTC)
You're telling me the best way to get Greek letters into an SVG image is to paste them in as graphics? That may be a workaround, but it's not really an easy one (I may see about converting the characters to paths in Inkscape), and it's definitely a bug. grendel|khan 17:53, 24 December 2006 (UTC)
The images on those pages are SVG already, so they don't have to be converted to paths. This workaround is a little cheesy, but at least the graphics you'll be "pasting" in are vectors. Mike Dillon 02:43, 25 December 2006 (UTC)
Moreover, if you get Greek letters working properly on Wikipedia, any user downloading the SVG picture with a bad local configuration won't be able to see the picture properly, while if you use the already-made SVG letters on Commons nobody will ever have any problem! Alessio Damato 19:44, 25 December 2006 (UTC)
It's not a bug. It's an incomplete implementation. The SVG backend we're using just doesn't support everything. For now, for maximum compatibility, make them paths. When support improves, you can revert to the old one using a character. —Simetrical (talk • contribs) 22:33, 25 December 2006 (UTC)
Wikipedia uses rsvg, right? I'm using librsvg 2.14.4, and the characters display fine in rsvg-view. Doesn't that make it a bug in Mediawiki? grendel|khan 18:41, 26 December 2006 (UTC)
No, it just means you have a font installed the server doesn't have. --cesarb 20:40, 26 December 2006 (UTC)
Ah, right, the servers badly need more fonts. There's probably a bug request somewhere, but if not, you could Mediazilla: open one. —Simetrical (talk • contribs) 22:31, 28 December 2006 (UTC)

Colored numbers in Watchlist

There used to be a huge mess of arguing and ranting here. View the page histories to find it; I thought it would be more productive to cut it down to this:

  1. The new numbers are a new feature which are the same as on recent changes. They refer to the number of bytes (roughly but not exactly, "characters") that have been added or removed from an article. A bug at the time meant that these couldn't appear when we introduced them for recent changes; we've fixed that now.
  2. If the numbers are showing up coloured for you, then this is because one or more of your local administrators caused this to happen; we did not set this.
  3. If you don't like it, then you can disable it, as described below.
  4. The notice on the watchlist page is nothing to do with us either.

To remove the byte counts, edit Special:Mypage/monobook.css or the appropriate skin subpage, and add the following lines:

.mw-plusminus-pos {display: none;}
.mw-plusminus-neg {display: none;}
.mw-plusminus-null {display: none;}

This will remove the display entirely, not just the colors. To remove just the colors, copy this but replace "display: none;" with "color: black;". To remove just the bold, replace "display: none;" with "font-weight: normal;". To remove the colors and the bold, replace "display: none;" with "color: black; font-weight: normal;".

Please remember that bug reports and feature requests need to be filed on our bug tracker at http://bugzilla.wikimedia.org.

Thank you for your co-operation.

Discussion

How does one do this without breaking the thing (i.e. just a bunch of randomly-formatted text appearing on monobook.css with no net effect)? This has happened every time I have tried to modify this page. Orderinchaos78 09:55, 24 December 2006 (UTC)
Purge your browser's cache. Titoxd(?!?) 09:56, 24 December 2006 (UTC)
Thanks for that - that seems to have got it to go in. However, the watchlist still displays as before. :( Orderinchaos78 10:00, 24 December 2006 (UTC)
Purge the watchlist as well... that should do it. Titoxd(?!?) 10:01, 24 December 2006 (UTC)
Ah yay! Gone finally! The first purge didn't work, but after a few minutes I went back again and it was exactly as it was this morning. Thanks for your help - much appreciated :) Orderinchaos78 10:09, 24 December 2006 (UTC)
I'll second (third, whatever) the opinion that in concept this was/is a great idea. And I'll also "second" the motion/wish that there are two lines added to preferences: percent value or byte value; and display or not (with "not" as the default). - jc37 09:29, 24 December 2006 (UTC)
Love the feature. Nice work. function msikma(user:UserPage, talk:TalkPage):Void 10:03, 24 December 2006 (UTC)

That software change was done by me to fix bug 8331. I'm sorry for doing something useful. --77.179.62.159 10:04, 24 December 2006 (UTC) (Leon)

No problem, thanks for your work. There are people that do like it :-) --Ligulem 10:37, 24 December 2006 (UTC)
Count me as another happy user; your hard work is valued. --mordicai. 18:26, 24 December 2006 (UTC)

Please remove the bold and color from the default and let the reader scan for what they want to know instead of competing with that activity. Why decide that for everyone? The colored number is unlikely to win a poll. I would bet the name of the article and user's name are used first. That said thanks for adding information to the summaries. -Susanlesch 10:04, 24 December 2006 (UTC) minor edit :10

The bold was already removed and the colors dampened. Force your browser to reload. —Centrxtalk • 10:09, 24 December 2006 (UTC)
Well done. Looks good. --Ligulem 10:37, 24 December 2006 (UTC)
Much better. :-) I wonder where all the comments went, at least I thought there were dozens about the bold. Anyway thank you. -Susanlesch 10:43, 24 December 2006 (UTC)

Is it possible to remove the notice (What do the coloured numbers mean?) and if so how. Harryboyles 10:16, 24 December 2006 (UTC)

  • I believe the coloured numbers refer to the number of bytes that have been added or removed from an article. (2nd sentence of #1 above) Keesiewonder 10:39, 24 December 2006 (UTC)
Add #watchlist-notice { display:none; } to your monobook.css ([9]) and force your browser to reload. --Ligulem 10:37, 24 December 2006 (UTC)

To just remove the bold, add the following lines to your monobook:

.mw-plusminus-pos {font-weight:normal}
.mw-plusminus-neg {font-weight:normal}
.mw-plusminus-null {font-weight:normal}

Hope that helps someone else; it was my only quibble and I saw it mentioned above :) Wonderful feature, much thanks. —Quiddity 11:04, 24 December 2006 (UTC)

For some reason, my post was removed, but I suggested an option to show percentage changes instead of bytes change. So blanking a short article might get -400 and a long article might get -10000. Either way, it will turn up as -99% or -100%. Insanephantom 13:19, 24 December 2006 (UTC)
Now my watchlist seems an IRC feed :-) Personally, I fully agree with Insanephantom: I would only find this useful if it told me the percentage that was removed (in example, -25% then I know someone removed a quarter of the article), and if it could detect null edits. Anyways, I guess some people will find this feature useful. Keep the good work up! -- ReyBrujo 14:07, 24 December 2006 (UTC)
The numbers are seemingly useful in detecting vandalism; I'm trying it on Recent Changes now and seeing how it works. However, like ReyBrujo said, displaying percentages might be more useful than displaying the number of bytes. -- Altiris Exeunt 14:11, 24 December 2006 (UTC)
I'm not seeing the colors at all in the Cologne Blue skin -- and I would like to see them. Also, I agree with the idea of showing percentages, but in addition to the bytes added/removed, something like "+250/+15%". Stevie is the man! TalkWork 14:24, 24 December 2006 (UTC)
I also noticed that some Wikipedias default to non-colored numbers (the Spanish and the Russian ones, in example). Another personal note: non colored numbers should be the default. But it is not a big thing, since we can modify that through css. -- ReyBrujo 14:31, 24 December 2006 (UTC)
Non-coloured numbers are the default in every Wikipedia unless the css has been modified. On the English Wikipedia, it has been modified to give the colours. Tra (Talk) 14:37, 24 December 2006 (UTC)
And the bold. By default, bold is reserved for edits that remove 500 or more characters. —Simetrical (talk • contribs) 14:56, 24 December 2006 (UTC)

I don't think the "what do the colored numbers mean" link should go here - this is too much about how the sausages are made, as it were. Perhaps there should be a simple explanation page in project space, with a link to this discussion. 90% of users who will click this won't want to be distracted by this discussion straight away. --ZimZalaBim (talk) 15:05, 24 December 2006 (UTC)

  • I don't see the addition as providing any meaningful information, since someone could make substantial additions and removals with a net effect of changing a small number of bytes. How do we remove this completely from the watchlist? --JJay 15:55, 24 December 2006 (UTC)
Edit your monobook.css, and change the new CSS classes to display:none (see above) // Laughing Man 16:02, 24 December 2006 (UTC)
Yes, that removes the colors, which makes it somewhat less distracting. But is there any way to remove this completely from the watchlist? --JJay 17:16, 24 December 2006 (UTC)
I like it! HighInBC (Need help? Ask me) 16:34, 24 December 2006 (UTC)
Sweet! I was a little confused at first, but nice addition! r9tgokunks 17:12, 24 December 2006 (UTC)
I don't know if this is just me, and I haven't touched anything, but the colour of the numbers has gone from normal red and green, to a dark red and green (like crimson and forest green). Disinclination 18:21, 24 December 2006 (UTC)
That has been done some hours ago, the other colors were very shiny. -- ReyBrujo 18:45, 24 December 2006 (UTC)
This is what I have done, it removed the message, unbolds it and reduces the size:
#watchlist-notice { display:none; }
.mw-plusminus-pos {font-weight:normal; font-size:75%;}
.mw-plusminus-neg {font-weight:normal; font-size:75%;}
.mw-plusminus-null {font-weight:normal; font-size:75%;}
HighInBC (Need help? Ask me) 18:33, 24 December 2006 (UTC)
Personally, I use { color:black; font-size: xx-small; }. -- ReyBrujo 18:45, 24 December 2006 (UTC)

Another way to remove the CSS styling is to use the special value "inherit" to use whatever the style would be without the overrides:

.mw-plusminus-pos,
.mw-plusminus-neg,
.mw-plusminus-null { font-weight: inherit; color: inherit; }

One problem that all of these CSS changes have is that the ".." after the diff numbers does not have a class on it, so it can't be modified or suppressed. Mike Dillon 18:56, 24 December 2006 (UTC)

I too would like those ..'s to be classed. HighInBC (Need help? Ask me) 19:04, 24 December 2006 (UTC)
For now to get rid of those ..'s you need to use this ugly javascript hack: User:GeorgeMoney/UserScripts/hacks/watchlist1. GeorgeMoney (talk) 04:27, 25 December 2006 (UTC)

Understanding

I wonder if i should make a note here but i had trouble understanding this from the notes above so this is briefly discussed on talk:Main Page. Just wondered if i should mention this. Simply south 14:42, 24 December 2006 (UTC)

An advantage i would say of these numbers is that you are able to tell more quickly who is blanking a page but then the trouble is that someone will beat you to reverting. Simply south 21:57, 24 December 2006 (UTC) :)

Login page and focus

I've been wanting to request this since my first login to Wikipedia. Incredibly I'm only doing it now and more incredibly it doesn't seem to be a FAQ (but I didn't search thoroughly): in the login form, could you please give focus to the username edit? A very small thing, but handy to have. —Gennaro Prota•Talk 15:26, 23 December 2006 (UTC)

Note that you should be able to put focus directly into the username field on the login form by hitting 'tab' once. --brion 01:30, 24 December 2006 (UTC)
Yes, I know, and that's what I'm currently used to do. But having a simple workaround doesn't mean not being a defect, no? :-) —Gennaro Prota•Talk 02:43, 24 December 2006 (UTC)
Lack of feature you want != defect ... besides, some people might prefer the focus set to the password field instead (for example, if the name field already has their username). And overusing .focus() is evil (tm). --Splarka (rant) 08:11, 24 December 2006 (UTC)
I thought the convention for a form was to set focus to the first input field. --CliffC 20:28, 24 December 2006 (UTC)
Indeed. But in the old times; now it's just a feature I [sorry for the markup abuse] want ;-) Admittedly I'm allergic to the mouse but anyway… Some people seem really to reply just for the sake of saying something. —Gennaro Prota•Talk 01:59, 26 December 2006 (UTC)

Edit count

How can I find out how many edits have I made ? Tavilis 08:25, 19 December 2006 (UTC)

Your edit count yandman 08:26, 19 December 2006 (UTC)

Thanks ! Tavilis 09:23, 19 December 2006 (UTC)


Goto the top of the page and click on my contributions pilko182 09:47, 19 December 2006 (UTC)

But that doesn't give you a count, just a list. User:Zoe|(talk) 18:40, 20 December 2006 (UTC)
Well, at least for a week or so more. It will probably give you your edit count soon now that there's a database field for it. —Simetrical (talk • contribs) 07:09, 21 December 2006 (UTC)
There's also a second problem with the my contributions page: viewing the oldest edits shows one fewer edit than it should. For example, when I click on the link for Oldest 20 edits, I'm shown this page, which actually only lists 19 edits. As it doesn't effect the other pages, I presume it's looking for a zeroth edit which doesn't exist. --Safalra 14:34, 27 December 2006 (UTC)

Fix for image links

Has the software been modified so that if certain explicit images are added to a page, they will not show up on the page as an image, but just as a link to the image. If so, then good. That will hopefully put a stop to stupid image vandalism.--Azer Red Si? 04:55, 23 December 2006 (UTC)

A long time ago. See MediaWiki:Bad image list. Titoxd(?!?) 08:17, 23 December 2006 (UTC)

SVG coming out blank

Image:Status_EPBC.svg does not render to PNG (it comes out all white). It's fine in Firefox (2.0.0.1 / Win) when you look at the SVG image itself. It was created in Inkscape, and various versions ("plain svg", converted to paths, inkscape original) are found in the edit history (detailed on the image page). Yes, I've tried purging. Any help? —Pengo talk · contribs 03:53, 23 December 2006 (UTC)

It works now. I started a new drawing in inkscape and copied the contents of the old one across. What exactly Wikipedia was choking on, I don't know. —Pengo talk · contribs 04:46, 23 December 2006 (UTC)

History RSS feeds not working?

My beloved RSS feeds aren't working anymore. I get the following error message in Firefox. It also doesn't work in Opera, so this doesn't seem to be a problem on my side.

XML Parsing Error: xml declaration not at start of external entity

Line Number 2, Column 1:
<?xml version="1.0" encoding="utf-8"?>
^

--Conti| 01:22, 23 December 2006 (UTC)

I'm getting an error as well. This could be because of a recent change to the code that caused them to become invalid. Tra (Talk) 01:37, 23 December 2006 (UTC)
This is caused by the xml declaration (<?xml version=.......?>) having a new line before it, so it is not on line 1 anymore and the xml parser goes crazy. This is probably due to some PHP include file closing and leaving 2 extra lines after the closing php tag (the first new line after a closing php tag is omitted). It could also be due to the server messing up the headers leaving one extra line. I have no idea. GeorgeMoney (talk) 03:34, 23 December 2006 (UTC)
Fixed. There was extra whitespace on the Special:Makebot extension's localization file. --brion 04:53, 23 December 2006 (UTC)

Refresh a Wikpedia page versus opening a new tab

I use Firefox, but don't think it matters to the question. If I have my watchlist loaded in a tab and refresh it, it takes many, many seconds to finish reloading. If I just open a new tab to load the page, however, it only takes a second. Why is there such a big difference? Thanks. Xiner (talk, email) 00:15, 23 December 2006 (UTC)

When you refresh, the browser has to ask the server, for each component of the page, if it has been modified since you last loaded the page. On the other hand, if you just open the page again (on a new tab or otherwise), it has to ask the server only about the components which have expired (or which were marked as not cacheable). The number of requests is much smaller on the second case; if you have a high latency connection to the servers, the difference can be quite noticeable.
My suggestion would be to simply load the page again instead of refreshing it (pressing enter on the URL bar is often enough). --cesarb 00:31, 23 December 2006 (UTC)

Pages only partially showing

I am only seeing Australian hardcore down to just above "See also", looking at it in edit the rest of the article is there. It is occuring in both Firefox and Explorer. Paul foord 23:45, 22 December 2006 (UTC)

It was a malformed <ref> tag; I've fixed it now. Tra (Talk) 23:57, 22 December 2006 (UTC)

Working redirect

The "comments" link in this talk template below is still red despite of existing redirect. Thoughts to fix it properly? --Brand спойт 21:21, 22 December 2006 (UTC)

I put ?action=purge on the end of the page URL to purge it, so it should be working now. Tra (Talk) 21:35, 22 December 2006 (UTC)

Zero sized reply from proxies when submitting

I've been getting this from show changes, submit, and preview actions. I've been mailing my edit to work and committing changes there without error, so I'm imagining that this is an issue with the proxy that serves my home IP address. The squid host in question is sq29.wikimedia.org. Is anyone aware of further details of this problem? Buffyg 19:58, 22 December 2006 (UTC)

I do sometimes find that none or part of a page loads. Try pressing refresh when that happens. Tra (Talk) 20:01, 22 December 2006 (UTC)
This isn't the same thing. It's not that the page doesn't load. It's that the proxy repeatedly gives an error message, and changes aren't committed. Reloading doesn't fix the problem. Buffyg 22:55, 22 December 2006 (UTC)
Try using the secure server (see the FAQ at the top of this page). This should bypass your proxy and workaround the problem. --cesarb 00:33, 23 December 2006 (UTC)

Image problem

The map on Long Island Rail Road is showing up way too large. No edits have been made to that article, template:infobox SG rail, or template:infobox rail that might have caused it. --NE2 23:18, 21 December 2006 (UTC)

I fixed it. The map size specified in {{infobox SG rail}} shouldn't have 'px' in it. Tra (Talk) 23:29, 21 December 2006 (UTC)
That's strange; the "300pxpx" worked until recently. I suspect whatever was changed broke other pages too. --NE2 23:31, 21 December 2006 (UTC)

Something is very odd. Up until yesterday List of largest suspension bridges looked just fine. Then suddenly, it looked like this! It was fixed by changing the parameter "150 px" to "150px". It doesn't seem that a space should matter. Worse, it seems like a bad idea to go from ignoring spaces to being fussy about them. -- Samuel Wantman 00:08, 22 December 2006 (UTC)

The devs know about it. Titoxd(?!?) 18:44, 22 December 2006 (UTC)
It's fixed in trunk and should be live sometime soon, if it's not already. For now, try to use correct syntax.  :) —Simetrical (talk • contribs) 19:11, 22 December 2006 (UTC)

What links here, my contribs, other lists?

The "What links here" list and the "my contributions" list both show a bulleted list of items. Other lists may have the same issue. To change it to a numbered list would require changing instances of <ul> to <ol> in the page, but seems it would be more useful and have zero impact on serving the page. Doesn't seem to be possible to change it with a user script - is it? If not, would it be a huge process to get this change implemented server side? Is it even a good idea? --*Spark* 22:41, 21 December 2006 (UTC)

The core question is "which item is marked as #1"? If you only want it to be the item at the top of each page, even as you hit (older 50) and (newer 50), then that can be easily done via monobook.js. However, that's not necessarily very accurate/consistent, so I don't know that should be presented to users by default. The most accurate would be to mark #1 as your very first edit, and have increasing numbers as you go up the page, and as you hit (newer 50). Then you wouldn't need edit counters anymore. :) From what I've been able to tell though, keeping track of that would be database intensive, [10] so it's not clear that solution could be implemented either. --Interiot 23:01, 21 December 2006 (UTC)
The only server side change I was referring to was changing the list element from unordered to ordered. Not true counting functionality on the server. Each page would start with #1. How would I do this with monobook.js? --*Spark* 23:15, 21 December 2006 (UTC)
In both cases, the <ul> doesn't have a class or ID, so you'd have to do some intelligent searching for ULs, I guess. --Interiot 00:56, 22 December 2006 (UTC)
Note that edit counts won't be so server intensive anymore, as a recent database schema change added an user_editcount field to the User table. Titoxd(?!?) 23:18, 21 December 2006 (UTC)
Cool. It still wouldn't necessarily help this case though, since you can stick random things in the from= field, and make Special:Contributions jump into the middle of your edits, and it wouldn't have any reference point as to the number of edits at that point. --Interiot 00:56, 22 December 2006 (UTC)
You can do this with CSS if your browser has decent support for it. If this list is numbered properly, it will work:
  • Foo
  • Bar
  • Baz
To get this, add body.page-Special_Whatlinkshere #bodyContent ul, body.page-Special_Contributions #bodyContent ul { list-style: decimal; } to Special:Mypage/monobook.css. Note that this will have to be updated with the appropriate page names if the base special page names are changed in the future (which they probably will be, to e.g. Special:What links here). —Simetrical (talk • contribs) 19:10, 22 December 2006 (UTC)

converting a jpg image to a url

please help me do this, so i can add a photo to a blog,

thanks so much and Happy Holidays,

Randi

Do you mean you want to add a picture on Wikipedia to your blog? If so, you will need to find the picture you want, right click it and select properties. Somewhere there, you should find a URL beginning with http://upload.wikimedia.org. This is the url for the image. If your blog gets a lot of traffic, it might be better to upload the image to another server to reduce the impact to Wikipedia's servers. Tra (Talk) 21:57, 21 December 2006 (UTC)
Not sure to understand the question clearly. Every file has an address, be it on your computer or on the web. A jpg image is a file ; its address is closely related to an url.
You are able to add a photo to a blog by stating its url, which you see in your browser when you go to that photo.
This is OK unless the image is used for a link, as our Wikipedia shattered globe top left does. It links to http://en.wikipedia.org/wiki/Main_Page, but its own address is ... http://upload.wikimedia.org/wikipedia/en/b/bc/Wiki.png. To taste the difference, copy each link in the address bar. Does it help ? -- DLL .. T 19:41, 22 December 2006 (UTC)

Bug/feature with your watchlist

The symptom: if you copy/paste a page off of your watchlist and use it to make a wikilink, it turns into a redlink. For example, Wikipedia:Administrators' noticeboard/Incidents‎ obviously exists, but is redlinked.

The cause: If you hit view->source while on your watchlist, you will see a non-printable character directly after the </a> and before the semicolon. This non-printable character gets included when you do a copy/paste from your watchlist.

The so what: On several occasions, I have seen pages that appear to have the correct name, but if you hover over the link, you see some junk at the end. If you hover over my link to WP:ANI above, you will see bad characters at the end. This can cause obvious problems. User A copies text off of his/her watchlist to use as a WikiLink in an article. Even though a real article exists, you now have a redlink making it look like no article is there. If someone clicks on it, they can create a new article that will appear to have the same name as the real one. You now have two forks of the article in existence. BigDT 21:48, 18 December 2006 (UTC)

It's a Unicode LRM. I just went to Special:Prefixindex and nuked six pages starting with it; unfortunately, there isn't a Special:Suffixindex to find the ones which end with it. I believe it shouldn't be possible to use it within page titles; I will file a bug asking for it to be removed from page titles. --cesarb 22:38, 18 December 2006 (UTC)
Reported as bugzilla:8312. --cesarb 22:50, 18 December 2006 (UTC)
Thank you for passing it on. I had seen this happen beore, but never knew where the incorrect character was coming from. It wasn't until this evening when I created the redirect WP:FURG by copying and pasting Wikipedia:Fair use rationale guideline from my watchlist (see [11] for that version) that I was able to figure it out. BigDT 22:54, 18 December 2006 (UTC)
Fixed by brion on r18513 as explained on his comment at bugzilla:3696. I already went through all the remaining articles with broken titles on this wiki and deleted or fixed them. --cesarb 13:08, 23 December 2006 (UTC)

SVG problem

If you look at Line (racing), the image currently doesn't have a red arrow on the pavement. If you look at commons:Image:Racing line.svg, the image does currently have the red arrow. There are two versions of the image in history (one with the arrow and one without), however the last modification was done 7 months ago, Special:Log also doesn't show any other activity at commons or here, and ?action=purge doesn't clear up the problem. Any idea what's causing this and/or the best way to fix it? --Interiot 22:50, 21 December 2006 (UTC)

I fixed it by reverting the last revision then reverting myself. Tra (Talk) 23:10, 21 December 2006 (UTC)
So do we need to do this every couple months, or is this something that can be triggered another way? (action=purge? Is there any other way to do a "null edit"?) --Interiot 00:50, 22 December 2006 (UTC)
The ways of fixing these kinds of problems I can think of are (in this order):
  • Refresh the page
  • Clear browser cache
  • Put ?action=purge on the URL of the image page
  • Put ?action=purge on the URL of the article page
  • Call the image with a dummy parameter in the querystring e.g. http://upload.wikimedia.org/wikipedia/commons/b/b7/Racing_line.svg?abc=def
  • Repeat all of the above several times over
  • Revert the image or upload another copy over the top (this would probably qualify as a null edit)
  • Make a null edit to the article
Or if all else fails:
  • Change the width parameter of the image by 1px
  • Create a new image from scratch
  • Upload the image under a different name and reference the new name
I'm guessing when/if this happens again, the best thing to do would be to try some of the things in this list and hope at least one of them works. Tra (Talk) 01:31, 22 December 2006 (UTC)

top of page links jump left

Tonight for some unknown reason all my user links at the very top of the page 'my talk' etc when I touch them with the mouse cursor they jump to the left and hide behind the jigsaw globe --Matt 10:37, 21 December 2006 (UTC)

I remember that happening to me as well on Internet Explorer. Have you tried Firefox? —Mets501 (talk) 12:04, 21 December 2006 (UTC)
Happens to me occasionally, in IE. DuncanHill 13:48, 21 December 2006 (UTC)
It has stopped again as mysteriously as it started, we have discussed firefox it might be time --Matt 21:33, 21 December 2006 (UTC)to look into it thanks
  • Bug Microsoft, it's a well-known problem on their end. Titoxd(?!?) 21:39, 21 December 2006 (UTC)

Problem on AfC pages

I'm not sure where this goes, so I'll start here.

For the last couple of days the AfC pages have been acting very screwy.

  • The "edit button" does not edit the line that it's on - it opens the article before!
  • The afc templates seem to be randomly putting the correct -top- and -bottom- tags on the new articles, so that sometimes editing a new article will affect the entire page!
  • When a new request comes in, ofttimes there is not a new == == heading created.
  • When there are multiple == == sections, attempting to edit the parent one opens only the top section, and not the entire submission. I've had to go through and blank out several nonsense listings that had multiple sections.

If this isn't the correct forum to address this, would you please redirect it to the right place! :) Thanks... SkierRMH 00:42, 21 December 2006 (UTC)

The last two aren't bugs. The third is just people not following instructions, and the fourth is a feature of section editing. -Amarkov blahedits 00:44, 21 December 2006 (UTC)
And the first?
Also, it is not consistent with the third one, sometimes all the sub-sections open up and sometimes they don't! SkierRMH 00:51, 21 December 2006 (UTC)
The first happens when you click it, and someone creates a new section above it. The sections are renumbered, so, for instance, if you click on the fourteenth, you go to what used to be the thirteenth. No idea about the second. And is there a reason the third can't just be people sometimes following directions and sometimes not? -Amarkov blahedits 00:56, 21 December 2006 (UTC)

Browser cache/vandalism question

I don't know what to make of this question about old vandalism showing in Safari; can anyone have a look and respond to this new user? Thanks, Sandy (Talk) 22:07, 20 December 2006 (UTC)

SVG rendering issue

See Talk:Keyboard layout#Standardisation_needed.

How can we purge the old thumbnails from the cache?

And how come it used to give a problem, but doesn't anymore? Shinobu 21:58, 20 December 2006 (UTC)

search spell correction

any reason we can't incorporate a spelling correction suggestion if misspellings are being search for within wikipedia? (i was thinking 'Did you mean xxxx?' google style). JoeSmack Talk 17:01, 20 December 2006 (UTC)

This feature has been disabled for performance reasons. Use Google to search Wikipedia instead. Tra (Talk) 18:32, 20 December 2006 (UTC)

javascript - those "are you sure you want to do that" with ok and cancel

hi i am researching javascript and need to find out the real term for those JS dialogs that say something like are you sure you want to do that, with ok carrying on, and cancel cancelling it? thanks —The preceding unsigned comment was added by 84.65.166.37 (talkcontribs) 22:26, 19 December 2006 (UTC).

var boolean = window.confirm("Are you sure?");

--howcheng {chat} 22:38, 19 December 2006 (UTC)

A confirmation dialog? Can't say that I've ever seen an official name for it. Cburnett 22:41, 19 December 2006 (UTC)
It would just be confirm(), not window.comfirm()
var boolean = confirm("Are you sure?");
GeorgeMoney (talk) 00:05, 20 December 2006 (UTC)
That's because if you call some function, the parser first looks for window.functionName() then functionName(). So window.confirm() is actually correct (it's a method of the window object) and also why you can just do confirm(). howcheng {chat} 01:04, 20 December 2006 (UTC)
Well i never knew that. But if it worked for you, then what was your original problem that was posted here? GeorgeMoney (talk) 03:21, 20 December 2006 (UTC)
I didn't ask the question, but it occurs to me now that I didn't answer it either. To answer, I don't think there's a name for it beyond "confirmation dialog box". howcheng {chat} 06:39, 20 December 2006 (UTC)
Oddly enough, W3 calls the three of them (alert, confirm, prompt) "popup boxes" [12], and the confirm() one a "Confirm Box". Netscape sometimes refers to them as a "browser dialog box". --Splarka (rant) 08:32, 20 December 2006 (UTC)
Windows programmers know these things as message boxes, regardless of the available buttons. The term message box is reserved for a dialog box containing only a message or question and a few buttons (and possibly an icon). But if the box asks a question, the term message box may sound a bit odd. It's a shame Javascript doesn't allow for other types of message boxes, like yes-no or yes-no-cancel. Ok-cancel message boxes really should only be used if you just asked the computer to do something, and you want the user to be able to cancel after reading the consequences. Shinobu 22:13, 20 December 2006 (UTC)

Fun new feature: edit change sizes in Recentchanges

After some database schema changes, we have started updating the live software on the site again. Visibly, this introduces a field on Special:Recentchanges showing the number of characters added or removed by each edit, giving an impression of the magnitude of changes.

This figure has been available for some time on the IRC feeds of changes, but is now stored in the database for display in the recent changes list. --brion 19:50, 19 December 2006 (UTC)

This is great! Would it be possible to add this to other 'lists of edits' e.g. watchlist, new pages etc. It might also be useful to add it to article histories and user contributions but this might not be possible if the field is stored only in the recent changes table. Tra (Talk) 21:52, 19 December 2006 (UTC)
One of the classes has a typo in it. class=\'mw-pluminus-neg\' in /includes/RecentChange.php line 565. (doesn't match the class "mw-plusminus-neg" in main.css) --Splarka (rant) 08:12, 20 December 2006 (UTC)
Fixed. —Simetrical (talk • contribs) 07:20, 21 December 2006 (UTC)

Automatic edit summaries

Can anyone remember where the discussions of the automatic edit summaries are taking place? The problem I want to raise concerns redirects. Currently, if someone creates a redirect from scratch, and saves the page without an edit summary, the automatic edit summary says <-Redirected page to XYZ, as seen here. But if someone then edits the redirect, adding a category for example (and there are reasons to add other text to redirect pages), but still leaving the page as a redirect, then removing all the stuff except the redirect produces the same automatic edit summary of <-Redirected page to XYZ, when in fact the redirect was in place all along. See the history here to see what I mean. Any way to get the automatic edit summary to only appear when a redirect is added and previously it wasn't there, and avoid doing this when both the before and after pages are the same redirect? Carcharoth 14:35, 19 December 2006 (UTC)

Specifically, this automatic edit summary is misleading. Carcharoth 14:36, 19 December 2006 (UTC)
Discussions are taking place at Wikipedia talk:Automatic edit summaries, where you could request this. Tra (Talk) 14:47, 19 December 2006 (UTC)
Thanks. Carcharoth 22:16, 20 December 2006 (UTC)

Special characters "edittools"

How do I get the special character box on my Media Wiki, or better yet, is there a help page on it? I've been looking for a while, and I can't find a help page on the special character box. My last resort was to copy the html... but I don't even know where the js function is. Could someone help me? By the way, I'm using media wiki 1.8.2. Hangfromthefloor 09:25, 17 December 2006 (UTC)

...never mind, I found MediaWiki:Edittools, but I still don't know how to use it. Also, can anyone give me suggestions for how to automatically change apostrophes and quotation marks in to smart ones? i.e. ' → ’ Hangfromthefloor 12:00, 17 December 2006 (UTC)

  1. You want the CharInsert extension. Check Special:Version for various things we have that you might be missing out on.
  2. You could write an extension to do a preg_replace on the page text on parse. The only tricky bit would be writing the regex so that it gets things that you want to get and not anything else. —Simetrical (talk • contribs) 07:12, 21 December 2006 (UTC)

new messages on talk page are non-existant

I keep getting a message that I have new messages on my talk page. However, when I check, I find no new messages. In fact, there have been no edits to that page in two days. What gives?

I don't know if it is relevant or not, but I use User:GeorgeMoney/UserScripts/newmessages. Should I ask him? Will (Talk - contribs) 04:19, 18 December 2006 (UTC)

But you have messages today. I posted a message to someone's talk page, and although it was listed in the user's talk page, the edit would not appear in the user's talk page history until a minute or so had passed. I believe there are some caching issues, so I would not really worry about this. -- ReyBrujo 05:34, 18 December 2006 (UTC)
Do you check your Talk page's history? Sometimes somebody might vandalize your page and another editor come along and remove it, thus triggering the "new messages" message. Or somebody could have added a comment and then thought better of it. Check the history of the page to see if anything like that occurred. If it's nothing like that, try editing the page to clear the message. User:Zoe|(talk) 18:39, 20 December 2006 (UTC)

Shortcut string bug

Firefox lets you make a shortcut string (like "wp") search Wikipedia and other sites. However, it turns '/' into "%2F". Wikipedia doesn't handle that correctly like it does when Firefox replaces ':' with a similar sequence. Is this fixable or do we have to wait for a MediaWiki update? Will (Talk - contribs) 06:11, 12 December 2006 (UTC)

In namespaces with subpages disabled, '%2F' should probably be accepted as an alias for '/'. You could open a bug report if you like. —Simetrical (talk • contribs) 01:53, 13 December 2006 (UTC)
This is a low-level issue with Apache; see my comments on the bug report. --brion 10:45, 14 December 2006 (UTC)
What specific bug report? Harryboyles 03:42, 21 December 2006 (UTC)
Bug 8254. —Simetrical (talk • contribs) 07:06, 21 December 2006 (UTC)

Undo

Undo is not undoing. It says it is, but the bad text remains in the edit screen. (It worked a couple of days ago, just not now - several times tonight.) Tvoz 08:28, 20 December 2006 (UTC)

Borders around thumbnail images gone

If this is intentional, I apologize profusely, but the pictures at Flag of Bulgaria look faintly ridiculous. It is interesting to note that this is not the case at Flag of the United States. Lockesdonkey 22:13, 19 December 2006 (UTC)

probablly a transiant screwup, i've just purged the page and it looks fine now. Plugwash 22:21, 19 December 2006 (UTC)
This was bugzilla:8326. Should be fixed now, though you might occasionally see badly mixed cached stylesheets. --brion 22:24, 19 December 2006 (UTC)

SVG arrow problems when rendering as PNG

New to creating SVGs with Inkscape, but have a look at Image:SPI single slave.svg and Image:SPI three slaves.svg:

 

and

 

When mediawiki is rending them as PNGs it must not understand the arrows on the lines and drops them. However, Seamonkey 1.0.6 and Firefox 1.5.0.8 render the SVGs fine (go to the image page then click on the file name underneath to see the actual image file). Anyone know how to get mediawiki to render those arrow heads? Cburnett 20:11, 19 December 2006 (UTC)

Yeah, you have to convert the lines to a path. Select the line, then choose Path > Stroke to Path. That should do it. howcheng {chat} 20:27, 19 December 2006 (UTC)
Every line I do that the arrow is no longer straight and at ~45 degree angle off of what it should be (both in inkscape and seamonkey). If it helps, I'm drawing a bezier curve with the line snapped straight by holding the control key, then end marker set in the fill and stroke dialog. Does it regardless of the line and marker type I choose. :/ Cburnett 20:35, 19 December 2006 (UTC)
Ok, it has something to do with the line being translated to a path resembling a filled rectangle to give it the appearance of being a line. And since the marker is anchored to a control point then, apparently, the angle of the triangle (the arrow head) depends upon the angle of the adjacent control points. Is there another way I could do this instead? Cburnett 20:50, 19 December 2006 (UTC)
Sorry, I only had to do it once, and it worked out fine for me, so I don't have any other suggestions for you. howcheng {chat} 21:10, 19 December 2006 (UTC)

Got it! I upgraded from 0.43 to 0.44.1 and repeated the strok to path operation. Seems ok. Cburnett 00:04, 20 December 2006 (UTC)

Image caption font size

 
This font size seems large.

The font size in image captions seems to have gotten large, but I can't figure out which CSS it's in. Does anyone know when/if this was changed? howcheng {chat} 19:52, 19 December 2006 (UTC)

Yea. I noticed it as well.. it is messed up, doesn't blend well at all. thanks/Fenton, Matthew Lexic Dark 52278 Alpha 771 20:14, 19 December 2006 (UTC)
I think it looks OK, now. Tra (Talk) 21:54, 19 December 2006 (UTC)
Probably related to bugzilla:8326. howcheng {chat} 22:37, 19 December 2006 (UTC)

Space in external links

A parameter of {{hpw}} allows for the user to insert the name of an article on the Harry Potter wiki, and it should link it to that article. However, if a space is used in the parameter (for example {{hpw|Harry Potter}}), the link will take you to http://harrypotter.wikia.com/Harry and pipe it to read "Potter". How can spaces be transformed to underscores in this parameter? --Fbv65edel / ☑t / ☛c || 18:17, 19 December 2006 (UTC)

Feed the parameter through the m:Help:Parser function urlencode (or anchorencode, see m:Help:Magic words). {{urlencode:Harry Potter}} results in Harry+Potter while {{anchorencode:Harry Potter}} results in Harry_Potter. -- Rick Block (talk) 18:37, 19 December 2006 (UTC)
Easiest would probably be: {{hpw|Harry_Potter|Harry Potter}} producing:
Be careful to avoid leading spaces on the first template parameter. Gimmetrow 18:39, 19 December 2006 (UTC)
Thanks, I just used {{anchorencode:{{{1}}}}} to avoid alerting people to use an underscore. --Fbv65edel / ☑t / ☛c || 19:15, 19 December 2006 (UTC)
{{anchorencode:}} is probably the wrong solution here: for example, {{anchorencode:M&M's}} produces "M&M's". That function is meant for MediaWiki section links, and isn't particularly useful for anything else. The correct solution in general would be {{urlencode:}} with the /w/index.php?title= link syntax; however, for Wikia wikis, an easier solution is to link to f.ex. wikia:harrypotter:Harry Potter. For other interwiki link prefixes, see meta:Interwiki map. —Ilmari Karonen (talk) 08:15, 20 December 2006 (UTC)

Page creation

how do you create a new page —The preceding unsigned comment was added by Pilko182 (talkcontribs).

Step 1: If you don't have an account, create one (see Help:Logging in). Rick Block (talk) 18:41, 19 December 2006 (UTC)
Step 2: Go to a relevant page, edit the existing page so the it will link also to the new page.
Step 3: In the edited page, follow the red link you just created, and it will create the new page
Step 4: Edit the new page. Thank you contributing wikipedia! Cate | Talk 10:11, 19 December 2006 (UTC)

What idiot starts from 0 when counting ;) 22:28, 19 December 2006 (UTC) :)

TPF

I want to see the PNRS that underwent schedule change on Nov 18. I work in Apollo test system. I understand that the PNRS will be logged into a queue. I want to know how to identify the Queue..Can any of you tell me the entry?

Not sure which part of Wikipedia this applies to. Can you narrow it down a bit? Notinasnaid 10:21, 19 December 2006 (UTC)
You might want try asking at the Reference desk. Tra (Talk) 14:49, 19 December 2006 (UTC)

Shrunken Text?

When I call up articles the text shrinks. I hate it! Why does it do this?

Text size shrinking upon log off

It's rather bizarre. All of a sudden, the size of body text--but not of tables of contents or text on the Main Page--shrink a size. I first noticed this at United States House of Representatives, accessed from the Main Page (in the Lewinsky Scandal bit). Is this on my side or yours? Lockesdonkey 03:12, 19 December 2006 (UTC)

You probably have a different skin set in your preferences. If you view it with the standard skin do you see a difference? --*Spark* 03:21, 19 December 2006 (UTC)
It's probably related to the issue brought up one section below this. Titoxd(?!?) 03:34, 19 December 2006 (UTC)

Special:Mostrevisions

What's the problem here? It's been offline for three weeks now.

It puts too much stress on the database servers; hence, it was removed. It hopefully will be brought back up when Wikimedia buys more servers, so donate! Titoxd(?!?) 21:44, 18 December 2006 (UTC)

Section headers on my monobook.js page don't work

At User:Will Pittenger/monobook.js, I configured it to make jumping directly to a specific script easier by adding section headers. The section headers are JS-commented out to prevent them from being executed. However, the page is shown in one big <pre> box when I view the page. How do I fix it so the section headers work again? -Will Pittenger 06:46, 12 December 2006 (UTC)

This has apparently been part of MediaWiki code for over 2.5 years: [13] (although until recently, it seems like it often failed to correctly escape all the html or apply the pre, if viewing a diff, old ID, or other strange circumstances, maybe something was fixed). What you can do if you just want to display (not edit), is transclude it to another page, eg: {{User:Will Pittenger/monobook.js}} on User:Will Pittenger/monobook. --Splarka (rant) 09:15, 12 December 2006 (UTC)
Having investigated further, it appears the only time you'll see your page in a slightly-wikicode-friendly manner is just after you've created it (and only you will see it as such, and only while it is cached, &action=purge turns it into a big escaped <pre>). --Splarka (rant) 00:28, 13 December 2006 (UTC)

Purge might work, but the purge tab does not. -Will Pittenger 01:25, 14 December 2006 (UTC)

Do you mean 'does not work' in that it makes it a big <pre> (which you don't want) or doesn't? --Splarka (rant) 09:29, 14 December 2006 (UTC)

The page is one large <pre> box until I purge the page. Same problem with my CSS page. Will (Talk - contribs) 22:13, 18 December 2006 (UTC)

Odd, I have the opposite effect. When I create a new css/js, it appears rendered in wikicode, but when I purge it, it henceforth only appears (as the code implies) in a large <pre>. --Splarka (rant) 08:08, 19 December 2006 (UTC)

Something wierd is going on....

I seems like the system isn't behaving normally. After making an edit and then looking at the history, sometimes my edit isn't listed. Other weird things happening that are hard to explain. Anyone else seeing things like this? ike9898 01:44, 18 December 2006 (UTC)

Oh, I see it is mentioned above. ike9898 01:46, 18 December 2006 (UTC)

My JavaScript isn't loading.. (possibly due to new "interact" sidebar)

My user scripts aren't loading; I think it's due to the fact that some admin made a new "interact" sidebar. --Split Infinity (talk) 00:36, 18 December 2006 (UTC)

There's quite a lot of scripts there. Try blanking the page then adding back each of the scripts one by one to see where the problem is. Tra (Talk) 00:42, 18 December 2006 (UTC)

I suspect the problem lies in the fact there was no longer a navigation box, I renamed the browse box back to navigation? —Ruud 01:05, 18 December 2006 (UTC)

That's right. Thanks! --Split Infinity (talk) 01:08, 18 December 2006 (UTC)

Edit article with a new section link

Is there a way in wikitext to add a link to edit an article?

Specifically, I want in my User talk page to have a link to edit my user talk page in a new section.

Is this possible, and if so then how.

Thanks. — Preceding unsigned comment added by xensyria (talkcontribs) 18:40, 17 December 2006

Sure, edit here. Gimmetrow 19:05, 17 December 2006 (UTC)
Thanks, that's very kind of you. I take it that there is no internal wiki way to do this?
Also, sorry about not signing it before. Xensyria 19:07, 17 December 2006 (UTC)
If by "internal" you mean as a [[wikilink]], then no, you cannot do this as this method requires parameters. However, you can make it a bit more classy: add new comment (as using the full url manually can break on mirrors and the https server). --Splarka (rant) 20:54, 17 December 2006 (UTC)

some developer has f***ed the database up!!

I saw someone delete my comment I made, but I don't think they were doing that, just trying to respond to it, but if you try to respond to it you end up deleting the message that was left! Anyone else got this problem? Paul Silverman 17:53, 17 December 2006 (UTC)

You mean by section editing? I could answer here properly. Prodego talk 18:21, 17 December 2006 (UTC)

Lost last half hour of edits

I was looking at Richard Dawkins, my watchlist had an edit from it recently: 17:39 Richard Dawkins (diff; hist) . . Lithfo (Talk | contribs), but when I go to history, the most recent edit shows as: 13:51, December 17, 2006 Mikker (Talk | contribs). Purge did not help. Seems all the articles on my watchlist only show edits up to 17:19 --*Spark* 17:49, 17 December 2006 (UTC)

This keeps happening. Sometimes the article history I'm looking at is current, sometimes it stop at around 17:20. A few refreshes helps, suppose I'm getting bounced around to different servers, some which are not up to date. --*Spark* 17:56, 17 December 2006 (UTC)
Similar thing here - a change shewed up on my watchlist, I went to look at it and then chased a websource, came back to wikipedia, edited the page, then clicked on my watchlist and was presented with an older version than I had been seeing 5 minutes earlier!DuncanHill 17:57, 17 December 2006 (UTC)
from observation It looks like there's a problem with Watchlists, but edits are being applied normally without disruption. --User:Ceyockey (talk to me) 18:03, 17 December 2006 (UTC)
Edit histories intermittently have the same problem, eg, an edit shews up on a users contributions page, you go to the article, the article shews the up to date version, but the last edit doesn't shew on the edit history.DuncanHill 18:07, 17 December 2006 (UTC)
same here Kotok-McCarthy. Susanlesch 18:08, 17 December 2006 (UTC)
  • The developers are aware of it, per #wikimedia-tech. Titoxd(?!?) 18:08, 17 December 2006 (UTC)
  • Wouldn't it make sense to lock the database in this sort of situation? --- RockMFR 01:51, 18 December 2006 (UTC)

edittools

I want to add EditTools to my MediaWiki 1.8.2 (you know, the character box on the edit page, etc.) except: 1) my wiki is kinda buggy, and 2) I have no clue where to get the javascript, and 3) I have no clue what the <charinsert> stuff is. Please help? 151.200.19.88 11:53, 17 December 2006 (UTC)

mw:Extension:CharInsert --Splarka (rant) 20:47, 17 December 2006 (UTC)

Firefox 2.0 refresh issue

Does anyone else have a problem in Firefox 2.0 where you edit an article, click Save Page, and the resulting view doesn't show your changes until you refresh the page? Just wondering if there is a workaround or fix. Mus Musculus 20:25, 15 December 2006 (UTC)

Yes, I have that problem as well. Prodego talk 22:30, 15 December 2006 (UTC)
It's probably that the squid cache hasn't been updated with the changes you made to the page. ?action=purge should deal with it. Titoxd(?!?) 22:35, 15 December 2006 (UTC)
I don't think it would be that, because the problem only occurs when the page edited is loaded immediately after the save, after which a soft refresh will load the page. Whereas if it were the squid cache, the page should continue to load the old version, until the squid cache is purged. Prodego talk 22:39, 15 December 2006 (UTC)
But there we go again: the squid cache doesn't take much to update, but it takes long enough for immediate views to be served using the old page. I have Firefox 2.0 as well, and I only see the problem when Ganglia says that the servers are overloading. Titoxd(?!?) 22:50, 15 December 2006 (UTC)
I have not checked that, however, I do notice the problem only occurs on long pages, so you are probably right, the squid did not update before the page was loaded, then when it is reloaded it has updated. I have no idea what you mean by "here we go again", has there been some debate on something like this? Prodego talk 23:00, 15 December 2006 (UTC)
No, just that I had written the same thing two lines above and in the Help Desk. :) Titoxd(?!?) 17:57, 17 December 2006 (UTC)
Well actually you may be right (unless this is a byproduct of the fact that all info is sent from or to the main server when this is selected), since going to my preferences and disabling page caching seems to have fixed it. Prodego talk 22:49, 15 December 2006 (UTC)

I'm having the same problem. It's only been happening over the last week or so for me. Firefox 2.0 on a Mac. I used to use the ?action=purge command each time, which sorts the problem on an individual page basis, but now I've given up and just assume that the edit has been saved. Mike Peel 20:52, 17 December 2006 (UTC)

Hmm - I wonder if that's the problem I'm having too: I posted this a few days ago on Help (didn't know about this place), and then the problem went away mostly, but it is still intermittently happening. I use Firefox too, but on a PC - didn't realize that might be relevant. Also, I don't know what you mean by "?action=purge" - please explain, if it's something I should be doing. However, this problem of not seeing the edit until I refresh the screen was not happening until a few days ago, and it doesn't happen each time, with no particular pattern. Went away for a few days, now is back, sometimes.
Here's what I said:
Reporting possible bug
Seems in the last few days whenever I complete an edit ("save" it I mean), I am returned to the page or section showing the old text in place until I refresh the screen, where it then shows me the updated text. This never used to happen (by "used to" I mean a couple of days ago) - previously after "save" I'd see the new text immediately. Is this a bug or feature, and if a feature, why? thanks Tvoz 18:40, 15 December 2006 (UTC)
Thanks for any help! Tvoz 03:25, 18 December 2006 (UTC)

Image Problems

File:SNCF Grenoble-Montmélian.svg
the image

Why does this image not show up inthumbnails (of frames, or in nothing). Chris5897 (T@£k) NOTE: The time, which is EIGHT O'CLOCK, FIFTEENTH OF DECEMBER, is left out to avoid automatic archival.

It references an external image, which a) doesn't exist and b) isn't allowed if it did. You need to either remove the image or tell your SVG editor to embed the images into the SVG file. --brion 20:32, 15 December 2006 (UTC)
Those GIF links are not the problem; they are not trying to include the GIF, and removing them makes no difference at all. In fact, the image as uploaded [14] works fine when viewed separately, only inlining it doesn't work (at least in Firefox 1.5.0.8). This isn't related to Wikipedia at all, hand-written HTML has the same problem. Looks like a Firefox bug. --Derlay 03:02, 16 December 2006 (UTC)
Hmm... It doesn't work on test.wikipedia.org, but it renders fine at Wikia, but we very recently (like, two days ago) upgraded our SVG rendering. --Splarka (rant) 08:38, 16 December 2006 (UTC)
Our SVG renderer is modified to fail automatically if the image attempts to load any external resources (for instance, external .gif images). Local files wouldn't work anyway, and allowing the renderer to load arbitrary files from our servers would be a security risk. Hence, failure to render. --brion 19:41, 17 December 2006 (UTC)

Problem of message alerts.

My message alerts don't go off, in spite of having read the message by clicking message alert link. They remain even after response for few days and always come on any page, I open. swadhyayee 13:30, 17 December 2006 (UTC)

Try refreshing, clearing your cache and purging both your talk page and the pages where the 'you have new mesages' alerts come up. Tra (Talk) 14:46, 17 December 2006 (UTC)

Thanks, but technically I am dumb to understand catch and purging so what are these two things and how to do it? swadhyayee 14:55, 17 December 2006 (UTC)

Oh, I am sorry, I have an article on my favourite list. I select the article from my favourite list after opening internet browser and I get the page with alert. To make it clear, I think this page is in my favourite list from long and message alert was not as default. swadhyayee 15:00, 17 December 2006 (UTC)

See Wikipedia:Purge for more information on how to purge and clear your cache. Thanks! Flcelloguy (A note?) 15:41, 17 December 2006 (UTC)

What's Going On?

About one out of every four times that I searched a keyword on Wikipedia in the past week, I got the little message that said that there was a problem with my search due to temporary technical difficulties. Is this problem the fault of my computer, or the Wikipedia server in general, and can I do anything about it? Eilicea 00:06, 17 December 2006 (UTC)

I think it's just a problem in general. You're not at fault, other users see it too. My guess is that it's a side effect of the servers having to try to index over a million articles and update the index every time someone edits, or something like that. What to do? Just try searching again in a little while, say 30 seconds. Nihiltres 01:06, 17 December 2006 (UTC)
At least for me, it's not been 30 seconds. In fact, it's lasted half the day on two occasions. I've been searching via Google, though--does that work for you?

Undeleting recreated articles

Why does it take so long for an article to appear (sometimes an hour or more) after an article that has been recreated is undeleted (i.e. the histories are merged)? —Doug Bell talk 21:26, 16 December 2006 (UTC)

It doesn't take any time for me, perhaps you need to purge the squid cache, or your own cache(Ctrl-F5 in many browsers, including IE and Firefox)? Prodego talk 21:29, 16 December 2006 (UTC)
I'm guessing it needed to be WP:PURGED. That seems to have fixed this, I'll remember that in the future if I have the problem again. Thanks, —Doug Bell talk 22:03, 16 December 2006 (UTC)

highlighting spaces?

Is it possible to highlight spaces in article differences? Many times I look at the diffrences and have to examine to find the difference. The Placebo Effect 02:58, 16 December 2006 (UTC)

Sure, there are a few ways to do it, which involve editing your user/monobook.css:
This method gives the differences a light-red background color, and forces all spaces to expand fully. Note that large edit pages may be hard to read with white-space:pre:
td.diff-deletedline span.diffchange {background-color: #ffdddd; white-space:pre }
td.diff-addedline span.diffchange {background-color: #ffdddd; white-space:pre }
If that looks too ugly, you can try something more subtle, like putting a padded border around the diffchange (which will be an empty box where spaces were added or removed, making them obvious) (although they will be collapsed):
td.diff-deletedline span.diffchange {border:1px solid #ff0000; padding:0 2px; }
td.diff-addedline span.diffchange {border:1px solid #ff0000; padding:0 2px; }

--Splarka (rant) 08:55, 16 December 2006 (UTC)

white-space: pre-wrap; is probably superior to white-space: pre;, if your browser supports it (can't recall if IE7 does). —Simetrical (talk • contribs) 05:32, 17 December 2006 (UTC)

fun and profit with templates

Hi. I've been trying to make, at User:Morwen/library, a table containing books. The idea is that I would feed in the {{cite book}} parameters, and it spits out a table row, which has the output of {{cite book}} in one column, and then the actual source code to invoke that in the middle. Thus far I have failed spectacularly. Any suggestions? Morwen - Talk 11:24, 14 December 2006 (UTC)

I believe this is impossible. Because if you write the curly template call braces in to the parameters of the call it is already too late. Because in

{{User:Morwen/metacite|1={{cite book|title=[[The Star Trek Encyclopedia]]|last=Okuda|first=Mike|coauthors=Denise, Okuda with Mirek, Debbie|publisher=Pocket Books|id=ISBN 0-671-53609-5|year=1999}}}}

I would say cite book is already expanded before User:Morwen/metacite is called. So I would say the parser gets:

{{User:Morwen/metacite|1=<cite class="book" style="font-style:normal" id="Reference-Okuda-1999">Okuda, Mike, Denise, Okuda with Mirek, Debbie (1999). <i>[[The Star Trek Encyclopedia]]</i>. Pocket Books. ISBN 0-671-53609-5.</cite>}}

Which is sure not what you want.
Furthermore, in your subtemplate, User:Morwen/metacite:
|-
|{{{1}}} || <nowiki>{{{1}}}</nowiki>
you are telling the parser with nowiki to ignore that {{{1}}} is a parameter. Impossibilities on two levels. --Ligulem 12:22, 14 December 2006 (UTC)
So do you have any idea how to accomplish this effect? I'm happy to mess around with stuff, I just want to only have the parameters in one place on the 'library' page, so as to avoid chance of mis-synching. I was working under the hope that it would do my cite book after it had sustituted in the parameters, and wanted it to nowiki after parsing the cite book, but before parsing the wikimarkup of that. Obviously it's not doing what I want. Morwen - Talk 12:24, 14 December 2006 (UTC)
"So do you have any idea how to accomplish this effect?": No. --Ligulem 12:45, 14 December 2006 (UTC)
I'm not quite certain I understand the confusion. Why not just have your metacite accept the cite parameters, then do a passthru of the parms. Here's a simple example of that, but note you'd have to add conditionals to test if parms are defined to show this cleaner, but I think the concept is there. See User:*Spark*/citeTest. --*Spark* 18:19, 14 December 2006 (UTC)
Hmm. Hmm, that has the drawback of not being very generic, but it will do for my purposes. Morwen - Talk 13:12, 15 December 2006 (UTC)
A book is a book is a book. The question is valuable, if the solution uneasy today. If a template is not made to be put inside another, there could be a way to link to some information from an URI, ISBN or Wikipedia page inside a template. Let's think about that. -- DLL .. T 18:56, 14 December 2006 (UTC)
Eh? Morwen - Talk 13:12, 15 December 2006 (UTC)

I think you can do something with msgnw. Put all the template calls in one subpage "cites". Then you can make a table with

{|
| {{User:Morwen/cites}}
| {{msgnw:User:Morwen/cites}}
|}

The first column will have all the evaluated templates from the cites page, the second will have just the code from the page. Gimmetrow 22:56, 16 December 2006 (UTC)

Get all page

On the dutch wikibooks I created an automatic referral system for a series of pages that form a book: at the end of every page a "turnpage"-template is added (without any arguments). The template contains information about the book (list of pages included in the book). The template automaticly generates a link to the "next" and "previous" page in the book, but of course the information in the template must be updated.

The template contains the following (input by user):

{{Turnpage|Name=Book

|1=Page1</nowiki>

|2=Page2

}}


(this template uses the template turnpage):

{{#if:{{{2|}}}|<span style="font-size:120%; color:blue; text-decoration:underline;">[[{{{1}}}/{{{2}}}|← Previous]]</span>}} • <span style="font-size:120%; color:blue; text-decoration:underline;">[[{{{1}}}]]</span> • {{#if:{{{3|}}}|<span style="font-size:120%; color:blue; text-decoration:underline;">[[{{{1}}}/{{{3}}}|Next →]]</span>}}

It is possible to generate this page automaticly?

  • How to 'read' all pages with the prefix Firefox (e.g.: Firefox/Installation, Firefox/Extensions)
  • How to input them in the template
  • If someone knows a better way to do it: tell me. 86.39.9.130 08:49, 16 December 2006 (UTC)
Just a thought: if you had all the turnable pages as sub pages in a subpage-enabled namespace, then you could use {{SUBPAGENAME}} sneekily: On /wiki/SomeBook/Page# you have two template inclusions like: [[{{BASEPAGENAME}}/{{next|{{SUBPAGENAME}}}}|Next →]], where Template:Next is: {{#switch:{{{1|}}}|Page1=Page2|Page2=Page3|Page3=Page4|...}} and Template:Prev is the reverse. You'd have to ask the devs to enable subpages though. --Splarka (rant) 09:12, 16 December 2006 (UTC)

After

After we log out when we are on an edit this page page, we should should go back to the page, with any edits we've made, but not submited yet, so we can do so. My user page is 100110100.70.74.35.164 02:16, 17 January 2007 (UTC)

getting content with python

I am trying to learn have to use some of python's HTML parsers and I thought it would be fun to work with wikipedia content. I tried to download a wiki page using the following line in a python script

f = urllib.urlopen("http://en.wikipedia.org/w/index.php?title=Pittsburgh%2C_Pennsylvania&oldid=101081233")

but I got the following error


Access Denied.

Access control configuration prevents your request from being allowed at this time. Please contact your service provider if you feel this is incorrect.


Is there a way to access (read-only) the raw html programmatically in the manner I am trying?

I'm pretty sure of that this is a case of stupid user-agent filtering (a vain attempt to fight spammers i presume). The solution which all spam bot authors alreayd know of is to simply change the user-agent string to something which doesn't begin with "urllib". Below is an example which fetches the page to a file.
import urllib

url = "http://en.wikipedia.org/w/index.php?title=Pittsburgh%2C_Pennsylvania&oldid=101081233"

class MyEvilBot(urllib.URLopener):
    version = "MyEvilBot/0.0.1"

urlopener = MyEvilBot()

urlopener.retrieve(url, "sumfile")
Jeltz talk 00:12, 17 January 2007 (UTC)

To add some explanation user agent is a something which the client sends to the server to indicate what program and version he is running. Since it is all up to the client to claim whatever he wish it is a very weak filter easily cricumvented which mostly just annoy legitimate users at no cost for the spammers. Jeltz talk 00:17, 17 January 2007 (UTC)

Hmm, realized that my posts must sound rude to the sysops who set up this filter. Sorry guys, didn't mean to sound rude but it came out that way. Measures against spammers which hurt legitimate users is a touchy subject for me. Jeltz talk 00:21, 17 January 2007 (UTC)

Thanks for the response, your solution worked. 71.61.219.140 01:29, 17 January 2007 (UTC)

I'm not sure what spammers you're talking about; the user-agent filtering is a speedbump to reduce the impact of poorly-written mass-download bots. You can easily set a useful user-agent string which identifies your bot, and always should. --brion 07:17, 17 January 2007 (UTC)
Ah, thanks for the explanation. It makes sense a lot of sense it is a good convention to have a user-agent string for your bot. I still don't like it but whatever, at least you have a good reason. :) Jeltz talk 12:14, 17 January 2007 (UTC)

References color of Harry Rimmer

I do not understand, but the first reference of Harry Rimmer seems to have a strange blue color. Did I do this? And how?--Filll 20:12, 16 January 2007 (UTC)

False alarm. I fixed it.--Filll 20:14, 16 January 2007 (UTC)

track facility

Hi. It would be great if one could have a tree type visual representation in the sidebar of the links one has visited - a sort of tree shaped history sidebar, with the option of allowing a person to save this history.

It's more likely that a feature like that would be implemented in your browser before it'll be implemented here. For example [15].
Alternatively, there's been a discussion of including typed links in each article, for browsers that support that, it's possible that might happen sooner. --Interiot 21:36, 16 January 2007 (UTC)

Thx for the links.

SQL error with the Query API

I'm getting an SQL error when retrieving revisions with the Query API, e.g. query.php?what=revisions&titles=Main%20Page&rvuniqusr&rvlimit=3&rvcontent. Might be the result of a recent database change? I only get the error when retrieving the revision content. If the 'rvcontent' parameter is removed it works. Cheers, Jayden54 14:05, 16 January 2007 (UTC)

You apparently also get it right if you remove "rvuniqusr" instead query.php?what=revisions&titles=Main%20Page&rvlimit=3&rvcontent. This may have been fixed by a very recent change to query.php [16] (not live yet). Tizio 17:49, 16 January 2007 (UTC)

Boldness degree of sections and subsections

Section headings, denoted by double =, and subsection headings, denoted by triple =, have their boldnesses backwards IMO. Some article authors, like those of Edgar Allan Poe, kludge their way out of this stylistic ugliness by skipping subsections altogether, going directly from sections to sub-subsections (quad =). I suggest the default CSS sheet be changed so as to make subsections have the boldness of current sections, and sections have the boldness of current subsections. Font height would remain the same (section headings slightly taller than subsections). I don't see how this would break existing article look and feel, and in fact would improve that of articles which already use sections and subsections. (I understand there is a way for a logged-in user to customize his own Wiki CSS, but for most users, and even editors, the default is what we see.) – 2*6 01:50, 16 January 2007 (UTC)

Actually, subsection headers have always been bold, but smaller in size than section headers (at least, that's how I see them in mozilla). I too prefer the old style. Tizio 17:51, 16 January 2007 (UTC)
It's true that first- and second-degree headings are not bold, because that generally looks ugly at their font sizes. This was a stylistic choice made by the author of Monobook and can be changed at MediaWiki:Monobook.css. It should not be changed on a per-article basis. If you would like to change it for yourself, you can add h1, h2 { font-weight: bold; } or h3, h4, h5, h6 { font-weight: normal; } to Special:Mypage/monobook.css. Note that boldness as typically implemented is not very granular: things are usually either bold or normal, you often don't have anything in between (although you can). —Simetrical (talk • contribs) 05:40, 17 January 2007 (UTC)
Thanks for the hints: I have added h3 { font-size: medium; } to my monobook.css, and everything look fine now. Apparently, the font size for headings are specified in monobook/main.css Tizio 13:01, 17 January 2007 (UTC)

Headers

I've got a problem with headers.

Basically, i've got an article where i need to use two levels of headers, and the second (lower) level needs the horizontal rule.

The problem is, only the = and == headers have the horizontal ruled line incorporated into them. Since the = header isn't really supposed to be used, we're stuck with using == and ===.

the === header doesn't have the horizontal ruled line incorporated into it, so we tried a ---- below the header. However, this creates a gap between the header text and the horizontal ruled line.

So the question is - does anyone know any tricks that i can use to get a horizontal ruled line incorporated into the === header (like the way the = and == header have lines on them). — Preceding unsigned comment added by Saintmagician (talkcontribs)

You could try putting in something like <h3 style="border-bottom:1px solid #aaa;">Insert text here</h3> but it is a bit of a hack and it probably won't look correct when looking at anything other than the default monobook skin on screen. Tra (Talk) 01:39, 16 January 2007 (UTC)
btw, why do you say this won't look right in anything other than the monobook skin? the h3 tag is html yeah? shouldn't it turn out the same no matter how someone's viewing it? --Saintmagician 02:02, 16 January 2007 (UTC)
One good way to test that out would be to play in the sandbox, or create your own sandbox (a subpage of your user page, e.g. User:Saintmagician/Sandbox) and try it there. You'd obviously need to reset your skin to view it in different ones, which is kind of a pain, but it's one way. --Tkynerd 02:13, 16 January 2007 (UTC)
It won't necessarily look "right" for anything other than Monobook because not all skins use the underlining and the border color or thickness may be wrong if they do. Mike Dillon 02:29, 16 January 2007 (UTC)
Why does it need the horizontal rule? Mike Dillon 01:44, 16 January 2007 (UTC)
Basically, these sections really should be == sections. It's for the digimon article merge that's happening right now. this was the original design (with sections 3-12 all used to be seperate articles, but now merged). However, it does look better to group those sections together as being 'other forms'.
So then it looked like this. With "other forms" being a == heading and the rest being a === heading. But we still wanted the horizontal rule to seperate the different forms. I thought this form looked bad, since there's a gap between the === heading and the horizontal rule we're using.
So i changed it to this - where i have the = and == headers, both with the horizontal rule.
But it seems like some people really don't like the = header in articles.
Just a thought, but does anyone here understand the technicalities of how the horizontal rule is incorporated into the = and == headers on wikipedia? --Saintmagician 01:57, 16 January 2007 (UTC)
This seems like a personal preference thing, not something needs to be done for everyone reading the article. As for the technicalities of the underlining of section headers, they are controlled by monobook/main.css (the first rule that concerns this starts with "h1, h2, h3, h4, h5, h6"). It is possible for people who want these sections to be underlined to do so in their personal CSS. I can show you how if you're interested. Mike Dillon 02:27, 16 January 2007 (UTC)
Mike, to clarify: You're talking about changing, in one's personal CSS, how the articles one sees will render, right? Not how the articles one edits will render for everyone. --Tkynerd 02:39, 16 January 2007 (UTC)
Correct. As far as I can tell, there is community consensus for the way headers currently display. If it really bothers someone that much, they can change the way they see it, but they shouldn't be changing the agreed-upon style of the site without getting some support from others. Style hacks don't help Wikipedia. Mike Dillon 02:42, 16 January 2007 (UTC)
Consider that my reply ("Rich's kludge"), offering a simple "solution" to the OP's question, seems to have been missed. Could my use of a "style hack" have contributed to this oversight? Probably. – 2*6 02:51, 16 January 2007 (UTC)

Rich's kludge


If a rule line-break, but not formal heading underline, would work, you could precede the text section with four dashes, as does this reply. – 2*6 01:58, 16 January 2007 (UTC)
Now! Even less kludgier!! Make a normal subsection declaration (with three =) and follow it on the next line with four dashes. Article text follows on the third line. This gives you a formal subsection and the underline. – 2*6 02:17, 16 January 2007 (UTC)

2*6 - regarding your suggestion. No, that's what i was trying to avoid. Using the '----' creates a gap around the horizontal rule and the text above it. Anyhow, i've ended up just using the horizontal rule above the headings. Thanks for all the suggestions never the less. --`/aksha 04:00, 16 January 2007 (UTC)

Maybe if the code .underlined-heading {border-bottom:1px solid #aaa;} is put into MediaWiki:Monobook.css and then the headings are written as <h3 class="underlined-heading">Insert text here</h3> then the underlines would only appear on the monobook skin and won't disrupt other skins. Tra (Talk) 16:20, 16 January 2007 (UTC)
Why should it be an editorial decision whether to add an underline to subsections? This doesn't seem to me like something that we should be promoting. That being said, you're solution is better than hardcoding the CSS style preferences into the page itself. Mike Dillon 16:30, 16 January 2007 (UTC)

blank pages?

I've noticed that sometimes (only started today I have no changed any settings), when I click to edit a page - it will appear as blank - even thought there is clearly content on the page - this occurs on both article and talkpages. --Larry laptop 22:04, 15 January 2007 (UTC)

I had this problem on pages that used the class="sortable" on tables. It turned out to be something in my monobook.js causing it. I blanked it to make sure and it fixed it. Then I had to restore one by one until I found out which was the problem. Hope this helps. --MECUtalk 23:15, 15 January 2007 (UTC)
I don't think this should happen anymore (at least not due to the sortable thing). The reason it happened is that in the Monobook skin the runOnloadHooks function is actually called twice, once from a <script> tag at the very end of the HTML content and once from the page's "onload" event. Inside runOnloadHooks, a variable called "doneOnloadHooks" is checked to see if the hooks registered with addOnloadHook should be run. This flag used to be set after all of the hooks had run successfully, but it is now set before any other code runs (thanks to a fix by Ilmari Karonen).
The upshot of this was that if a hook failed previous to the change, doneOnloadHooks would not be set to true and the hooks that had already run would execute twice. This includes the code that installs the JavaScript handlers for sortable tables, since it happens before any hooks added with addOnloadHook. I believe the reason that the page blanks is because the sortable table code uses document.write to include "sortable.js", which doesn't work when it is attempted from an onload hook. It should probably be changed to use something like importScript, which is found in our MediaWiki:Common.js.
I'm not sure why Larry's page is blanking though. Mike Dillon 00:10, 16 January 2007 (UTC)
The whole sorttable.js setup is somewhat hacky. I'm starting to think it might be better to just merge the code into wikibits.js. I'm not even convinced there's any significant performance gain from only including it on some pages — it might even hurt performance more than it helps, since loading it takes an extra HTTP request with all the latency that implies. —Ilmari Karonen (talk) 04:29, 17 January 2007 (UTC)
Could be, yeah. That might be sensible. —Simetrical (talk • contribs) 05:27, 17 January 2007 (UTC)