Wikipedia:Village pump (technical)/Archive 39

From Wikipedia, the free encyclopedia

[edit] blockquote

<blockquote></blockquote> does not indent when an image is to the left of the blockquote using <blockquote></blockquote>. Could someone please post this on bugzilla or whatever? Thanks!68.148.164.166 (talk) 08:26, 17 May 2008 (UTC)

This is how HTML is supposed to behave. The images only push the text aside as much as they need in order to display the image. If the width of the image is larger then the width of the blockquote indent, then this indent might be "non visible". --TheDJ (talk • contribs) 15:44, 17 May 2008 (UTC)

Yes. If you need to move the text around like that consider using one of the quotation templates such as {{quote box2}}. --— Gadget850 (Ed)talk 15:54, 17 May 2008 (UTC)

{{quote}} probably mimics <blockquote> better. Gary King (talk) 19:19, 17 May 2008 (UTC)

True, since it uses <blockquote>, but it give no ability for size, margins or alignment. ----— Gadget850 (Ed) ^talk - 22:51, 17 May 2008 (UTC)

[edit] Where are the Mediawiki foundation servers located?

Can any one point me to a map or list or something? I am curious because I never get lag or denial of service so I figure I must be close. - Icewedge (talk) 18:29, 17 May 2008 (UTC)

m:Wikimedia_servers. Equazcion •✗/C • 18:37, 17 May 2008 (UTC)

Not much there on locations actually. They're in the state of Florida, but I don't know anything more specific than that. Equazcion •✗/C • 18:39, 17 May 2008 (UTC)

The primary databases are in Florida. In addition, caches for the purposes of reading pages and delivering images are located in Amsterdam and Seoul. Dragons flight (talk) 18:43, 17 May 2008 (UTC)

Ok, Thanks. It might be the Seoul server that is me fast access. - Icewedge (talk) 18:49, 17 May 2008 (UTC)

[edit] Untraceable margin

Anybody has an idea why {{Parish church}} gets a big top-margin when used in St Patrick's Church, Hove? Circeus (talk) 01:46, 18 May 2008 (UTC)

It's to accommodate the size of the image. Try paring the image down to 200px or so. Cheers! —Ashanda (talk) 01:50, 18 May 2008 (UTC)

Even though I've looked at the template code and know there is patently no way this could be the solution, I tried. Lo and behold, it doesn't work. Circeus (talk) 03:32, 18 May 2008 (UTC)

Fixed. --- RockMFR 06:17, 18 May 2008 (UTC)

[edit] change the SEARCH field input box, please

Please make the SEARCH field input box *bigger*, *more prominent*, and at the *top center* of the home page.

Searching for articles is what people come to Wikipedia *for*. Making the search box small and sticking it in an obscure corner of the page is odd, awkward, silly, and for some visitors (elderly, vision impaired, or otherwise physically or mentally challenged) troublesome enough to make them give up and look somewhere else for their information. —Preceding unsigned comment added by 192.159.110.7 (talk) 00:12, 18 May 2008 (UTC)

S/he may have a point there. Perhaps the search box would be more visible to new users if it were directly under the puzzle globe? —Ashanda (talk) 00:21, 18 May 2008 (UTC)

Directly under the globe is exactly the right place. Someone looking for an article wants to go directly to searching for it, not scan down through two boxes and 11 links. Someone interested in finding out more about Wikipedia can easily skip the search box and browse through the links below it. -- John Broughton (♫♫) 01:34, 18 May 2008 (UTC)

The Commons' way of doing it isn't bad either. Instead of a list of portals or categories, they just put the search box right at the top of the page. —Remember the dot ^(talk) 01:53, 18 May 2008 (UTC)

Quick mock-up. I think it would be nice to integrate a search box into the Main Page somehow. Thoughts? AmiDaniel (talk) 03:30, 19 May 2008 (UTC)

[edit] Adding Applet to my user page

Hi All

I am a registered user but not in the English Wikipedia.

Can I add an applet to me user home page at wiki? and if so I can I do it?

Thank u in advance —Preceding unsigned comment added by 89.139.159.30 (talk) 06:10, 18 May 2008 (UTC)

If you are asking about another language version Wikipedia, the answer is "no" - Wikipedia is not Facebook. And adding applets raises security issues. -- John Broughton (♫♫) 15:55, 18 May 2008 (UTC)

[edit] Wikipedia as a chat client?

I just got this rather odd edit to my user talk page.[1] It occurred to me that while running Huggle I've sometimes seen what look like chat conversations going on as vandalisms to different articles. I wonder, is it possible that someone's using the recent changes feed as some sort of chat client? —Elipongo (Talk contribs) 16:11, 18 May 2008 (UTC)

Didn't we have a long discussion recently on this? In any case, given how quickly most vandalism is reverted, and how quickly an account can be blocked for repeated vandalism, it seems to me that Wikipedia is spectacularly inefficient as a chat room, and that the Internet has plenty of free alternatives far better suited for this type of thing. -- John Broughton (♫♫) 16:41, 18 May 2008 (UTC)

[edit] User talk:Neil- the new talk page and my sig are allergic to each other

Hi my sig works fine as far as I know on every other page. But when it's put on User:Neil's it turns bits of his text there green afterwards. I have cut and pasted the code he gave me, which just meant two square brackets showed after my sig, could you look at the sig and the page concerned, and see what's happening? Because until then I can't post to the lovely Neil's page. :( Anyway, my sig is orange, not green, so I'm completely at a loss and intrigued as to why this would happen. Hope you can help.Sticky Parkin 22:08, 18 May 2008 (UTC)

Your sig does not terminate it's tags properly; html tags started inside a wikilike must be closed inside the wikilink (not outside). I've corrected you sig above (see diff). — Edokter • Talk • 22:22, 18 May 2008 (UTC)

aaah, thanks. I think my clipboard was clogged too, so when I copied a bit of code it still was pasting an old version in to my preferences. I had to try turning it off and on again. :) Sticky Parkin 23:09, 18 May 2008 (UTC)

[edit] Sandbox

Isn't working. —Preceding unsigned comment added by 76.194.245.224 (talk) 22:30, 18 May 2008 (UTC)

It was unreasonably large for a while. It's been cut down to size now, which seems to fix it. Algebraist 22:39, 18 May 2008 (UTC)

[edit] Links to disambiguation pages (Warning when linking to...)

I spend a fair amount of time repairing links to disambiguation pages. It has struck me that if editors received some kind of warning when introducing such links then they would be more likely to realize their mistake and correct it themselves. Would it be possible for some such to be introduced? DuncanHill (talk) 22:32, 15 May 2008 (UTC)

Great idea but perhaps a technical hurdle. Even better would be showing them a list of options and automatically piping the link when they select the right one. xenocidic ( talk ¿ review ) 22:37, 15 May 2008 (UTC)

Yes, that sounds even better! DuncanHill (talk) 22:38, 15 May 2008 (UTC)

That could be a WP:WIKED enhancement but I don't think the standard Wikipedia editing interface can do that, ever. Pegasus «C¦T» 03:33, 16 May 2008 (UTC)

It could, sure. Wouldn't be too hard to do if it were desirable. —Simetrical (talk • contribs) 17:48, 16 May 2008 (UTC)

I've made a user script: User:Splarka/dabfinder.js to find disambiguations on a given page. However, it will not work correctly in preview because it utilizes generator=links on a page to find all template links on pages linked to from a given page and see which match the list defined at MediaWiki:Disambiguationspage (in this way, it can check all the links on a page with only two API queries). But it is a bit useful after the fact. --Splarka (rant) 07:14, 16 May 2008 (UTC)

Nifty. If you don't mind searching out all links in the document and doing ceil(N/50)+1 API queries instead of only 2, you can do something like I do in User:Anomie/linkclassifier.js:

1. Use getElementsByTagName('A') to find all links
2. Filter out links with classes 'external', 'extiw', 'new', or 'image'.
3. List the unique title attributes; this attribute contains the target page name, with a few exceptions (filter out '' and anything beginning 'Edit section:' or 'Special:'). Alternatively, you could decode the page name out of the href.
4. If you wanted, you could filter out anything with a namespace to be left with just mainspace page links.
5. Instead of using generator=links, pass up to 50 of the linked pages at a time in the titles parameter.

Interesting approach, I didn't know there was an official list of disambiguation templates. My script looks for categories the page belongs to to do the same thing. Anomie⚔ 11:23, 16 May 2008 (UTC)

That's what the official list does too, more or less. Actually it differs slightly, but probably on Wikipedia they amount to the same thing. —Simetrical (talk • contribs) 17:48, 16 May 2008 (UTC)

Some disambiguation pages are a symptom of bad planning and I would be wary of software features that try to discourage the user from linking to them. Smoking, Acoustic guitar, Hip-hop are just a few examples of proper articles which started out as disambiguation pages (for decidedly unambiguous terms) linking to intimately related sub-concepts at the expense of the bigger picture. — CharlotteWebb 20:53, 19 May 2008 (UTC)

[edit] Any special settings were necessary to make Template:Navbox work?

I am attempting to port Template:Navbox and it's variants to another wiki. Initially I ported the Templates source directly to a matching article on the other wiki. Instead of the normal Navbox rendered information, it returned a great deal of HTML and messed with the style of the page putting text behind the wiki logo, pushing the tool/link boxes on the left all the way to the bottom, and it even spilled the /doc page text out of the bottom of the normal article area of the page. In an effort to fix this I began playing around with it, enlisting the aid of User:CapitalR on a separate wiki environment that I used for a while to house information on custom content in an online game. Either way, CapitalR stripped out a lot of the parser functions and other stuff, and simplified it some, and it still seems to be having some difficulty. If you would like to take a look at the output of the full template as it is ported you can look here [http://www.mendo.org/ultima/Template:NavboxTest]. What I really need to know is, where there any special settings that had to be made to allow the template (in its current form or similiar previous ones) to function. Barring that, is there any useful help that anyone might give on it.

Already suggested/tried are the following:

• Common.css and Common.js sections that are used by this template may not be present - They are
• HTMLTidy may be interfering - HTMLTidy was not initially enabled so far I could tell. I have added some lines to the LocalSettings.php file to try and enable it just to see if it would fix it. and it did not change at all.

$wgUseTidy = True;
$wgTidyConf = "$IP/includes/tidy.conf";
$wgAlwaysUseTidy=true;

• Parser Functions are out of date - While they were pretty out of date, they are not up to date with version 1.1.1 displayed (although the files are reported as 1.12...

Any thoughts? Tigey (talk) 01:41, 16 May 2008 (UTC)

You might be missing an extension that's necessary: I can't off the top of my head think which it might be. What extensions are listed on en.wiki's Special:Version#Installed extensions list that isn't on the equivalent page on your wiki? Happy‑melon 09:06, 16 May 2008 (UTC)

OK there are quite a few of them, but most of them I can't see affecting it, then again, you never know...

All that are on my wiki are:

• Confirm User Accounts
• CharInsert
• Inputbox
• ParserFunctions 1.1.1

Perhaps tonight I will go through and install some of the likely extensions and see what I can make of that. Tigey (talk) 13:12, 16 May 2008 (UTC)

Of Note! I installed ExpandTemplates and it outputted the code like so:

<table cellspacing="0" class="navbox nowraplinks"><tr><th style="" colspan=2 class="navbox-title"><span style="font-size:110%;">Title</span></th></tr>
<tr style="height:2px;"><td></td></tr><tr><td class="navbox-abovebelow" style=";" colspan="2">Above</td></tr><tr style="height:2px;"><td></td></tr><tr><td class="navbox-group" style=";;">Group1</td><td style="text-align:left;border-left:2px solid #fdfdfd;width:100%;padding:0px;;;" class="navbox-list navbox-odd">List1</td></tr><tr style="height:2px"><td></td></tr><tr><td class="navbox-group" style=";;">Group2</td><td style="text-align:left;border-left:2px solid #fdfdfd;width:100%;padding:0px;;;" class="navbox-list navbox-even">List2</td></tr><tr style="height:2px;"><td></td></tr><tr><td class="navbox-abovebelow" style=";" colspan="2">Below</td></tr></table>

The output on the NavboxTester page displays the Title cell below this output:

<tr style="height:2px;"><td></td></tr><tr><td class="navbox-abovebelow" style=";" colspan="2">Above</td></tr><tr style="height:2px;"><td></td></tr><tr><td class="navbox-group" style=";;">Group1</td><td style="text-align:left;border-left:2px solid #fdfdfd;width:100%;padding:0px;;;" class="navbox-list navbox-odd">List1</td></tr><tr style="height:2px"><td></td></tr><tr><td class="navbox-group" style=";;">Group2</td><td style="text-align:left;border-left:2px solid #fdfdfd;width:100%;padding:0px;;;" class="navbox-list navbox-even">List2</td></tr><tr style="height:2px;"><td></td></tr><tr><td class="navbox-abovebelow" style=";" colspan="2">Below</td></tr> 

So it processes part of it correctly. Starting with that "Above" cell all the way up to the end table tag. Tigey (talk) 22:02, 16 May 2008 (UTC)

Looks like the necessary fix was an upgrade of the MediaWiki itself as suggested at Template talk:Navbox#Ported to outside Wiki - Problem if no image defined. Tigey (talk) 17:38, 19 May 2008 (UTC)

[edit] Adding article sizes in histories

Is there any way that the article's character count can be added next to each listing in the article's history? If I look at Recent changes, it shows how many characters have been added or removed with an edit, but this information doesn't show up in the article history, so there's no way to tell which edits did damage to the article. Corvus cornix_talk 18:34, 16 May 2008 (UTC)

Er... it does for me :D Happy‑melon 18:41, 16 May 2008 (UTC)

It wasn't this morning. Maybe I needed lunch.  :) Corvus cornix_talk 20:52, 16 May 2008 (UTC)

You can get the +/- difference for each edit in the history tab by subtracting byte counts, if that's what you meant:

 addOnloadHook(function(){
   if(wgAction != "history") return;  var b, b1;
   s = getElementsByClassName(document, "span", "history-size");
   for(var i = 0; i < s.length; i++, b1 = b){
      if(m = s[i].innerHTML.match(/[\d\,]+/)) b = parseInt(String(m).replace(/\,/g, ""));
      else if(s[i].innerHTML.match(/empty/)) b = 0;
      else return; if(!i) continue; d = b1-b; x = String(d);
      while (x.match(/\d{4}/)) x = x.replace(/(\d{3})(?:\,|$)/, ",$1");
      f = Math.abs(d) > 1000 ? " style=\"font-weight:bold;\"" : "";
      if(d<0) c = "neg"; else if(d==0) c = "null";  else {c = "pos"; x = "+" + x; }
      s[i-1].innerHTML += "</span> <span class=\"mw-plusminus-" + c + "\"" + f + ">(" + x + ")";
      }
   });

— CharlotteWebb 16:25, 19 May 2008 (UTC)

[edit] Okay I need a bit of help here with some formatting and stuff....

Okay, so I need some help... it's kind of hard to explain what I'm trying to do but... I'm trying to have it so that when you visit a page (my userpage, specifically) it will be different every time, like there will be random options for words.... I'm not sure how to achieve this.... I've seen on other wikia's where they have template:verb or template:noun or something like that, and they have it all formatted with different options of verbs and nouns and such. I would just steal the formatting from those pages but pretty much all other wikias are blocked on this computer :( anyways does anyone know how to make this work? let me know at my talk page please. -Guitarplayer001 —Preceding comment was added at 22:05, 18 May 2008 (UTC)

You want {{rand}}. Gary King (talk) 16:18, 19 May 2008 (UTC)

[edit] Infobox titles centered with the [hide]/[show] button?

Is there any technical way to have the title of the infobox centered without moving it to another line or adding any space characters in the front of the title? Because there is a [hide]/[show] button the title cannot be centered. And it doesn't look so good. I hope there is a way to "fool" the software to make the title centered.

I know that you can add "state = plain" attribute so there won't be a hide/show button, but I mean for those infoboxes that use the button. ---Majestic- (talk) 23:08, 18 May 2008 (UTC)

Add {{pad|5.7em}} before "Reviews". For a more elegant solution, see the code for {{navbox}} that just resolved this issue. ----— Gadget850 (Ed) ^talk - 02:25, 19 May 2008 (UTC)

I can't find it, as Navbox is a very complex template. Please show me the code. ---Majestic- (talk) 07:26, 19 May 2008 (UTC)

Just because I know it is in there does not mean I know how it works. I took a quick look, but I don't see how it works either. ----— Gadget850 (Ed) ^talk - 08:36, 19 May 2008 (UTC)

I will try {{pad|5.7em}}. There's one problem though. When one adds the "state = plain" attribute the table will always be expanded and the [hide]/[show] link on the right will not be displayed. But with the {{pad|5.7em}} the title is no longer centered either. Can you add an "if" parameter so that if the "state = plain" is used the "{{pad|5.7em}}" will not be used? ---Majestic- (talk) 09:05, 19 May 2008 (UTC)

I added {{#ifeq:{{{state|}}}|plain||{{pad|5.7em}}}} before "Reviews" and it seems to work. ---Majestic- (talk) 10:02, 19 May 2008 (UTC)

[edit] Image page file histories

Has someone been changing the format of file histories on image pages? My image tagging script (Howcheng's) isn't working anymore. Kelly ^hi! 02:03, 19 May 2008 (UTC)

Yes. --brion (talk) 02:08, 19 May 2008 (UTC)

Why would this be changed without an announcement? I imagine this is going to cause problems with image-processing bots as well. Kelly ^hi! 02:38, 20 May 2008 (UTC)

Screen-scrapping bots can expect things to break from time to time. Bots and JS should use the API to avoid that. Aaron Schulz 03:53, 20 May 2008 (UTC)

[edit] "Unauthorized" page content for anon users looks wrong

After visiting

http://en.wikipedia.org/wiki/Image:Oliver_Sacks.jpg

I tried to visit

http://en.wikipedia.org/w/index.php?title=Image:Oliver_Sacks.jpg&action=edit

as an anonymous user. The page title is "Unauthorized", but below that, a box says, "Wikipedia does not have a Image page with this exact title. ... Log in or create an account to start the Image:Oliver Sacks.jpg page".

Seems it should be saying that the page exists but that I, as an anonymous user, can't edit it. In any case, I could use some kind of change to the message; "Unauthorized" and "page doesn't exist" in the same page leave me confused. —Preceding unsigned comment added by 68.125.166.190 (talk) 00:59, 20 May 2008 (UTC)

The image is uploaded on Wikimedia Commons, meaning the image shows up, but the page does not exist. The reason it is unauthorized is because IPs can't make new articles, which you are actually doing. If you want to edit, make an account at Commons and change it there. Soxred93 (u t) 02:12, 20 May 2008 (UTC)

[edit] Is there a way to count how many articles are in a given category and ALL of its subcategories?

Other than manually :) I would like to know how many Poland-related articles do we have, which should be possible to find out if we could count articles from Category:Poland and its subcategories. Currently my estimate is based on roughly counting stubs, comparing them to project tagged stubs, and using the ratio to estimate all article. Such an estimation, of course, is far from satisfactory (FYI, it gave me a numbers: ~300 tagged stubs, ~9000 stubs total, ~3000 tagged articles, ~100,000 Poland-related articles total). The 100,000 number itself seems more or less correct, but I'd like a more precise estimate - since the error margin here is quite big.--_{Piotr Konieczny aka Prokonsul Piotrus| talk} 13:53, 18 May 2008 (UTC)

The editor's index reveals that User:Chris G Bot 2 does this. Algebraist 14:10, 18 May 2008 (UTC)

When I need a quick count I usually fire up AWB and just click "make list", but I expect it would time out with such a massive number of articles to assemble. Plus there's the usual problem of Wikipedia's category tree being sufficiently jumbled that you're not going to get a very accurate count for "poland-related articles" from this analysis anyway: how closely related do you think Scripps National Spelling Bee is to Poland? (That's Category:Poland → Category:History of Poland → Category:Military history of Poland → Category:Military operations involving Poland → Category:Wars involving Poland → Category:World War II → Category:Events cancelled due to World War II, if you want to look it up. Category:World War II is a favourite of mine in pointing out things like this, as it seems to have worked itself into some of the most unlikely parent categories: I found it in a distant subcat of Category:Thailand once!). How accurate do you want to be? Happy‑melon 14:11, 18 May 2008 (UTC)

OT:It's at Category:Thailand → Category:History of Thailand → Category:Wars involving Thailand → Category:World War II. That's not very distant. Algebraist 14:48, 18 May 2008 (UTC)

I once determined you can get to something like 80% of Wikipedia through subcats of Category:Sports. As I recall finding the path from Sports to Religion was one of the more complicated. Dragons flight (talk) 16:22, 19 May 2008 (UTC)

Gah, I haven't thought of that. Yes, that kills that idea right there :> --_{Piotr Konieczny aka Prokonsul Piotrus| talk} 14:28, 18 May 2008 (UTC)

Perhaps WikiProject Poland should consider creating something like the List of mathematics articles, though that takes a fair amount of effort to create and maintain. Algebraist 14:20, 18 May 2008 (UTC)

Some time ago I prodded List of Poland-related topics as it was indeed not maintained. Plus, can you imagine a list with 100,000 entries? There is Category:Poland-related lists and forgotten Wikipedia:WikiProject Lists of basic topics/Draft/List of basic Poland topics. No, as far as those lists go I'll stick to supporting only the assessment-related Category:Poland-related articles by quality, but as I pointed out above - those still cover less than 10% (probably around 3%) of eligible topics.--_{Piotr Konieczny aka Prokonsul Piotrus| talk} 14:28, 18 May 2008 (UTC)

We at Maths survive OK with a list with 20,000 entries, and I suspect that 100,000 is a substantial overestimate: it seems likely that non-stubs are much more likely to get tagged than stubs, and unlikely that as many as 4% of all en.pedia articles are Poland-related. Algebraist 14:43, 18 May 2008 (UTC)

That's an interesting hypothesis - I would love to test it (see how large a % of en wiki is related to pl subjects), but I cannot think of any feasible methodology to do it (at least not anything short of many hour effort, which may be sensible if I were to write a paper based on a results, which I am not planning to do - yet... :).--_{Piotr Konieczny aka Prokonsul Piotrus| talk} 18:14, 18 May 2008 (UTC)

Looking at the index, perhaps User:Erwin85/CatCount is a better place to get a count; as I read it, User:Chris G Bot 2 will produce an actual listing, which doesn't seem to be what is requested here. -- John Broughton (♫♫) 16:37, 18 May 2008 (UTC)

As far as I can tell from the documentation, CatCount doesn't delve into subcategories. Algebraist 17:02, 18 May 2008 (UTC)

If I give you a list of subcats would you be willing to filter out the non-pl cats? β^{_{command 2}} 16:37, 19 May 2008 (UTC)

This sounds feasible. If we can figure out a way how to get the Polish data, we could streamline the process and generate some useful results for other wiki areas :) --_{Piotr Konieczny aka Prokonsul Piotrus| talk} 13:27, 20 May 2008 (UTC)

[edit] Can we produce a time graph / stats of article's activity?

Is it possible to feed an article (talk page, etc.) into some bot or other tool, to get information on how many edits (and editors) contributed to each over time (for example by month)? PS. I am relatively familiar with editor-side counters, but not with article-side counters, which is what I need for this question.--_{Piotr Konieczny aka Prokonsul Piotrus| talk} 18:10, 18 May 2008 (UTC)

That's what wikidashboard's for. Algebraist 18:42, 18 May 2008 (UTC)

Thanks, but that tiny graph at the top, with no extractable data, is more of an eye candy than any serious research tool :( --_{Piotr Konieczny aka Prokonsul Piotrus| talk} 22:34, 18 May 2008 (UTC)

In that case, you probably want Wikipedia page history statistics, or one of the other options listed in the index. Algebraist 06:59, 19 May 2008 (UTC)

Ah, those look much more promising, thank you.--_{Piotr Konieczny aka Prokonsul Piotrus| talk} 13:29, 20 May 2008 (UTC)

[edit] Pages renamed

How and why are the pages listed as renamed on these pages being renamed? Wikipedia:Version 1.0 Editorial Team/DC Comics articles by quality log and Wikipedia:Version 1.0 Editorial Team/Marvel Comics articles by quality log. Thanks for any help, Hiding T 18:30, 18 May 2008 (UTC)

They're pages that have been moved between runs of User:WP 1.0 bot - when it doesn't find a page that it found last time, but it finds a 'new' page with similar content, it twigs to the fact that it's been renamed and logs this appropriately. Happy‑melon 18:56, 18 May 2008 (UTC)

But there's nothing in the move logs, or deletion logs. Not that I can see, anyway. Hiding T 19:17, 18 May 2008 (UTC)

For example, ah, they've been moved back. But Sigmar (Marvel Comics) (edit|talk|history|links|watch|logs) shows no record of being moved to Sigmar (Marvel comics) or even deleted, even though the bot log states it was renamed from Sigmar (Marvel Comics) to Sigmar (Marvel comics). What's the bot picking up? Hiding T 19:23, 18 May 2008 (UTC)

There appears to be a relevant log entry. —Simetrical (talk • contribs) 14:06, 19 May 2008 (UTC)

Relevant, yes. An explanation, no. That log entry is the exact opposite of what Hiding is talking about. We're changing comics to Comics, not Comics to comics, and yet the pages Hiding refers were listed as being renamed from Comics to comics, which is wrong. Doczilla _STOMP! 08:13, 20 May 2008 (UTC)

I help maintain the WP 1.0 bot. I ran it again on each of the two projects above, and the May 20th log entries seem to show the bot picked up the changes but didn't notice all them were moves. The way that the bot detects changes is somewhat fragile and changes in mediawiki must have broken it. I have enabled extra debugging output to help figure out what's going on. If (when) additional errors like this occur, please contact me directly. In the meantime, I will put it on my list to reimplement the code that determines the new name of a moved article. — Carl (CBM · talk) 10:58, 20 May 2008 (UTC)

I see. Looking here, it suggests that

• Sigmar (Marvel Comics) (talk) Stub-Class (No-Class) renamed to Sigmar (Marvel comics)

(so "Comics" to "comics"), which is the opposite of what the history says. Well, the bot is at least consistent with itself in the quality page (see towards the bottom).

In this particular case, the bot may have been confused by the rename of the quality page, see the history. This is the first time people noticed this, and from the code it is not clear what causes this. Let's see if it happens again apart from quality page renames. Oleg Alexandrov (talk) 03:53, 21 May 2008 (UTC)

• In this particular case the bot would also be confused by the fact that the move happened after it logged it. We moved it to where it should be after the bot told us it had moved, but we still can't track down when it was originally moved. At this point this thread is getting confused, so I will post a timeline to clear up some confusion:

1. 17th May, the bot records Sigmar (Marvel Comics) (talk) Stub-Class (No-Class) renamed to Sigmar (Marvel comics) [2]
2. 18th May, J Greb moves Sigmar (Marvel comics) to Sigmar (Marvel Comics) [3]

• The bot can't have seen that move as it hadn't happened yet, and it only happened because the bot saw a move we can't find. If the bot looked into the future and saw a move we hadn't yet performed and recorded it wrongly, causing us to move the page, that would be a temporal paradox and a physics paper and potential nobel prizes all round. So at this point we're still all perplexed as tpo why the bot thought the page had moved, yes? Hiding T 17:19, 23 May 2008 (UTC)

[edit] Special:Undelete text

What is "Undeletion will not be performed if it will result in the top page or file revision being partially deleted. In such cases, you must uncheck or unhide the newest deleted revision." talking about when you go to undelete a page? It doesn't make any sense - how can you partially delete a file or revision? I would remove it or ask on the appropriate talk page, but I can't even find where it is coming from - it is not in MediaWiki:Undeletehistory. --B (talk) 21:08, 19 May 2008 (UTC)

This is the best place to ask - MediaWiki talk pages like this aren't well-watched. It's in MediaWiki:Undeleterevdel which doesn't have any history because it was never changed from its default value. As to what it's about ... I have no idea ... maybe it refers to changes in revision compression over the years? The message must have been created after MediaWiki default deleted messages that were the same as their default value in January 2007, because it has no deleted history. Graham87 01:30, 20 May 2008 (UTC)

Hrm ... no wonder I couldn't find it - I was trying to search for the text, which doesn't exist. ;) Can anyone see a really good reason not to blank this message? It doesn't make a bit of sense. --B (talk) 03:21, 20 May 2008 (UTC)

I'm guessing that that must be a warning that a revision will not be undeleted if it is a revision that contains only metadata and does not reference an entry in the text table (e.g. after moving or protecting a page). That said, it is only a guess. Titoxd^{(?!? - cool stuff)} 03:30, 20 May 2008 (UTC)

Is that something that would at all be applicable on English Wikipedia? I googled "MediaWiki:Undeleterevdel" and found a number of other Wikis that have one more sentence added — "Revisions of files that you don't have permission to view will not be restored." That makes sense. If your access level doesn't give you rights to the file itself and you try to restore the file, nothing will happen. That could be true somewhere ... but not here as everyone with access rights to restore a file has access rights to view that file. The other thing I thought about that it could mean, but unfortunately doesn't, is that you shouldn't restore just a file but not the image description page. Unfortunately, though, it lets you do that. That's a rather bad thing, IMO, for obvious reasons. So if anything, there should be a warning not to do that, but the message as it is now doesn't seem to apply here. --B (talk) 03:38, 20 May 2008 (UTC)

(unindent) For future reference, Special:AllMessages lists all of the system messages and their default content and current content. A quick search of the page usually turns up what you're after, in this case, MediaWiki:Undeleterevdel. --MZMcBride (talk) 03:45, 20 May 2008 (UTC)

The message pertains to Special:RevisionDelete. Aaron Schulz 03:52, 20 May 2008 (UTC)

What is Special:RevisionDelete? I'm guessing it allows you to delete one revision rather than the whole article? If so, why don't we have that? It would be great to be able to take out one revision of libel rather than having to do a selective delete/restore or get an oversight user to handle it. --B (talk) 03:56, 20 May 2008 (UTC)

its still in developement. β^{_{command 2}} 14:32, 20 May 2008 (UTC)

In the interim, is there any reason that we can't blank MediaWiki:Undeleterevdel? It just makes no sense whatsoever when you go to undelete a page. --B (talk) 14:36, 20 May 2008 (UTC)

Well, if we blank it, it would show the default message hard-coded into MediaWiki. However, that default message is what we're seeing right now. A better way to fix the problem is understanding what the message is supposed to say and copyedit it. Titoxd^{(?!? - cool stuff)} 20:30, 21 May 2008 (UTC)

Makes sense. What's the status of Bug 3576? Is it still pending review? Titoxd^{(?!? - cool stuff)} 20:30, 21 May 2008 (UTC)

Special:Revisiondelete is not in development. It's finished. — Werdna talk 16:05, 20 May 2008 (UTC)

Is it going to be implemented here? Is there any reason not to blank MediaWiki:Undeleterevdel or replace it with something useful? --B (talk) 17:11, 20 May 2008 (UTC)

The page doesn't exist currently. If you "blanked" it, it would revert to the default message. --MZMcBride (talk) 19:44, 20 May 2008 (UTC)

[edit] How to add the logo in media wiki

HI All

I dont know how to add the logo into media wiki.

Can any one help me out in this regard. —Preceding unsigned comment added by 61.95.163.2 (talk) 14:17, 20 May 2008 (UTC)

If you have access to LocalSettings.php on your site, you can change the path/url value of the $wgLogo variable. Failing that you can use javascript or css to overwrite the background-image setting of the link (<a> tag pointing to your site's main page) within the div with id="p-logo". — CharlotteWebb 13:49, 21 May 2008 (UTC)

[edit] search for red links

Is it possible to do a search for non-existent pages that none the less are being linked to (matching some pattern?). Regards, /Marmelad (talk) 16:32, 20 May 2008 (UTC)

Use Special:WhatLinksHere/This page doesn't exist, replacing "This page doesn't exist" (which, surprisingly, is being linked to) with whatever you're looking for. Special:WhatLinksHere, predictably, also works for pages that exist. x42bn6 Talk Mess 16:48, 20 May 2008 (UTC)

If you want a database of all links, however, it's not available although you can run queries on database dumps, for example, WP:WANTED uses a dump from September 2007. x42bn6 Talk Mess 16:51, 20 May 2008 (UTC)

Special:Wantedpages is theoretically available, but it's disabled here, probably because it tries to sort the pages by number of incoming links or something equally horrible and expensive. If you had a more precise request, someone with a data dump or toolserver access could give you want you need, probably. —Simetrical (talk • contribs) 14:09, 21 May 2008 (UTC)

[edit] Autoconfirm now requires 10 edits.

Per Wikipedia:Autoconfirmed Proposal/Poll, which showed 85% consensus of 100-200 editors in raising the autoconfirmed requirements to ten edits, this has been implemented. — Werdna talk 00:03, 21 May 2008 (UTC)

Perhaps it would be appropriate for an announcement of the closing of the poll, preferably accompanied by a (brief) rationale for the decision taken, to be posted in the relevant discussion page? After all, it might not be intuitive to all why, with the majority supporting the seven-day, twenty-edit option, a lower limit was implemented. Waltham, The Duke of 00:51, 21 May 2008 (UTC)

Yes, what happened with that? The responses seem to be overwhelmingly in favour of the 7/20 option, by a very wide margin. (There are far more people in support of 7/20 then there were for all of the other options combined; 92 for 7/20 versus 58 for all of the others.) --Ckatz^chat_spy 05:16, 21 May 2008 (UTC)

Support was 85% for the implemented option, and 65% for the 7/20 option. Systems administrators have learned the hard way that 65% is not enough on enwiki. — Werdna talk 05:31, 21 May 2008 (UTC)

That assumes that everyone who voted for the 7/20 would also be happy with 4/10 and that no one who voted for 4/10 would be happy with 7/20. Both of those suppositions are dubious. Dragons flight (talk) 05:39, 21 May 2008 (UTC)

Oh good. Now Grawp will have to make ten typo fixes or space insertions before moving pages. Our problems have been solved!!!!!!!! --- RockMFR 19:33, 21 May 2008 (UTC)

Well, just watch for new accounts that are doing nothing but inserting meaningless spaces. It'll certainly slow him down. —Remember the dot ^(talk) 03:06, 23 May 2008 (UTC)

[edit] Talk:Giovanni Rucellai

What is causing that page to appear at Category:Candidates for speedy deletion. Thanks. CambridgeBayWeather Have a gorilla 05:17, 21 May 2008 (UTC)

[edit] Table formatting

How do I remove these border lines from this table? --soum ^talk 05:27, 21 May 2008 (UTC)

I just took care of it for you. --CapitalR (talk) 05:39, 21 May 2008 (UTC)

Thanks a ton. :-) --soum ^talk 05:41, 21 May 2008 (UTC)

[edit] Volapük Wiki

The Volapük wiki is currently listed as one of our major projects on the main page, and their stats page also has some interesting numbers, all of which seem to be ridiculously out of proportion to the figures at the Volapük article here, listing an estimated 20-30 speakers. One or the other is wrong, or that language has some bloody hard-core speakers... +Hexagon1 ^(t) 08:49, 21 May 2008 (UTC)

Load vo:Special:Random a few times; let me know how many clicks you managed before you A) got bored, or B) found an article that wasn't about a geographical location. The Volapük version of Blofeld of SPECTRE has been very busy over there... Happy‑melon 09:38, 21 May 2008 (UTC)

Most of the Volapük articles appear to be bot created. It's prominent placement in language lists has been discussed in different places. See for example Template talk:Wikipedialang#Quality requirements (Volapük). PrimeHunter (talk) 11:51, 21 May 2008 (UTC)

Ah, OK. That's somewhat beyond a joke, but it looks like it is still wanted. +Hexagon1 ^(t) 12:30, 21 May 2008 (UTC)

Took me 74 clicks before I got something, a date: vo:Novul 14, and 114 clicks to get to an actual article: vo:Heyuannia. Prodego ^talk 21:42, 21 May 2008 (UTC)

[edit] Search For User Contributions On A Specific Talk Page, Article Page, Category Page, Etc. Etc.

I proposal a way to search for all the contributions by a the same contributor on a specific talk page, article page, category page, etc. etc.

We can go to User contributions and see what the user contributed, using filters and using the drop down box to filterout for say contributions in the Articlespace. We need to search for contributions by editors for contributions in an article space for a specific article, let's say, so we can track down all the vandalism to the article contributed by the editor, let's say that it was pov. Please post this on the proposal media bugzilla or whatever cause I don't have an account. Thanks so much!68.148.164.166 (talk) 20:18, 21 May 2008 (UTC)

This should be perfectly doable now that we have a page_user_timestamp index on the revision table. —Simetrical (talk • contribs) 14:58, 22 May 2008 (UTC)

See this over at persistent proposals. If it doesn't incorporate completely what you propose, feel free to add. Impin | {talk - contribs} 17:54, 22 May 2008 (UTC)

its possible to do if you know how to use the API. β^{_{command 2}} 22:33, 22 May 2008 (UTC)

[edit] Age calculation

Is there any tag which calculates the age if we supply the birthday. I think this would really make it simpler on many articles where otherwise someone would have to edit the age all the time. Ninadhardikar (talk) 11:04, 22 May 2008 (UTC)

{{age}}. Anomie⚔ 11:09, 22 May 2008 (UTC)

I often see {{Birth date and age}}. Nihiltres^{t.l} 14:26, 22 May 2008 (UTC)

[edit] Information box

Could someone with technical knowledge create a header for this that says: "Legislative Assembly of New Brunswick" similar to the box below for Federal members of Parliament with a gray background header that says: Parliament of Canada:

Tank you. Jonathan Logan (talk) 14:03, 22 May 2008 (UTC)

Use {{s-par|ca-nb}} like this:

--—— Gadget850 (Ed) ^talk - 14:20, 22 May 2008 (UTC)

For the formatting of the box itself, I highly suggest, per succession box standardisation guidelines, to use spaced en dashes for years, as in 1900 – 1901. Waltham, The Duke of 23:27, 22 May 2008 (UTC)

Missed that bit. --—— Gadget850 (Ed) ^talk - 00:19, 23 May 2008 (UTC)

[edit] Strange block bug

User:N9NE Group was username blocked, then unblocked to allow namechange. The user now reports that he can't submit a namechange request, as the software is still treating him as blocked. Has anyone encountered this problem before? Algebraist 22:25, 22 May 2008 (UTC)

might me an autoblock. β^{_{command 2}} 22:30, 22 May 2008 (UTC)

[edit] Section edit link problems with Firefox

Resolved.

I am working to get List of Edmonton Oilers players to feature list status, and one of the feedback comments was that the section edit links are "pushed to the bottom of the article". Further investigation revealed that this happens when using Firefox, but I don't see it when I use IE7. The section edit links are rendered just to the left of the images, as expected. (I am unable to install a new browser on the PC I am using to verify this for myself.) This list has nine images as thumbnails attached to the top of the page, and for aesthetic reasons when rendering on different screen sizes, I'd like to keep them there. I'm looking for help from anyone familiar with Firefox who can verify what is happening on this list and perhaps find a solution that fixes the problem for that browser. Thanks — Andrwsc (talk · contribs) 22:52, 22 May 2008 (UTC)

To resolve edit bunching, see WP:BUNCH. Can be fixed with {{FixBunching}}. --—— Gadget850 (Ed) ^talk - 22:54, 22 May 2008 (UTC)

Excellent! Thanks for the link; I had never seen that phenomenon nor that template before. — Andrwsc (talk · contribs) 23:07, 22 May 2008 (UTC)

[edit] Duplicate userids?

Is it possible for two users to have the same userid? I've noticed before that someone seems to be making edits, even reversions, with my userid.

I haven't put much on my watchlist (not needed to: I've only once been reverted but that was a contentious and anonymous edit); nonetheless, maybe half the stuff on my watchlist was put there by someone else.

I've made, I dunno, of the order of 100-200 revisions (I can't work out the relationship between what you see and what you request from earlier times), mostly minor, and I don't THINK this other guy's stuff appears there. But he/she is surely getting his stuff on my watchlist page.

So could there be two of us with the same userid? Edetic (talk) 02:49, 23 May 2008 (UTC)

Can you provide a few diffs of the other user? Nakon 03:13, 23 May 2008 (UTC)

Page moves can add to your watchlist, check the logs of some pages you added to the watchlist, to see if they were ever moved to titles of pages you didn't add. And are you sure you don't have "Add pages I edit to my watchlist" ticked in your prefs? --Splarka (rant) 07:21, 23 May 2008 (UTC)

It's not possible for two people to have the same user ID, no. It would be possible, but exceptionally unlikely, that someone else is somehow being accidentally registered as editing under your account due to a Wikipedia error. Could you give more details on why you're so sure that you didn't watch these pages? Have there been any edits under your name that you're certain you didn't make? Do you have a plausible hypothesis for why anyone would watch all these pages when they're not editing them? —Simetrical (talk • contribs) 18:20, 23 May 2008 (UTC)

Hang on guys, red face dept here. I think maybe I've misunderstood the format of stuff that appears on a watchlist. Please relax (pending my finding time to look more closely), and let me say thanks for your quick responses. Edetic (talk) 18:37, 23 May 2008 (UTC)

[edit] Message for blocked IPs is defective

If someone edits as an anon from a blocked IP address, he/she gets the message "You are currently unable to edit pages on Wikipedia..." etc. However, further down it says "...please have your teacher or network administrator contact us (with reference to the following IP address: Crewe and Nantwich) at unblock-en-l...". In my case, I was trying to edit the Crewe and Nantwich article, and what seems to be happening is that the page's name is being output instead of the user's IP address. Can someone fix this? Thanks.--Anthony Nolan O'Nymous (talk) 11:33, 23 May 2008 (UTC)

Thanks for reporting, there was made a change to the schoolblock template which caused this. I've restored the previous version until someone has a better idea. --Oxymoron⁸³ 12:02, 23 May 2008 (UTC)

Thanks.--Anthony Nolan O'Nymous (talk) 12:13, 23 May 2008 (UTC)

[edit] Very long pages

Hi there, I'm currently going through the bot approvals process at Wikipedia:Bots/Requests_for_approval/FritzpollBot to create a bot which will complete populated settlement coverage within Wikipedia to stub level. The process has two steps, the second of which is to actually create the articles. The first step, however, places a processed list of the names, district, longitude, latitude on a Wikiproject subpage (for example, the bot created the list at Wikipedia:WikiProject Missing encyclopedic articles/Places/afghanistan - restricted to 100 edits) A change has been made to the bot to make it output the entire list onto such a subpage in one go, rather than one item at a time.

My concern is actually looking ahead to, say China, which has 87,000 different places. If I list these all on the same page in this format, I'm going to easily beat 1MByte, and there are a few other pages like this. I'm wondering if this will have any technical implications, and how I should mitigate it? (more pages with a certain limit?) Thought I'd check before I break Wikipedia! Fritzpoll (talk) 13:45, 23 May 2008 (UTC)

Any bot script that uses Wikipedia as an extended memory bank strikes me as rather inefficient, to say the least... is there any reason why the page would need to be available for other wikipedians to edit? If not, what possible benefit is there to posting it to wiki rather than keeping it as a data file on your computer/server? Happy‑melon 13:56, 23 May 2008 (UTC)

Yes, that would be inefficient. The reason (as perhaps not brilliantly explained on the approvals page) is that it allows human correction of article names (for example, where a topic with the name of a location exists, but has nothing to do with that location) and checking of district names to make sure that the internal links within the article are correct. What I could do is only output the fields needing checking to the page, and use the wiki-page to index a local file with the full info in it. I'd have to make absolutely certain the order hadn't been changed within the wiki-page to prevent screwing the process up, but it's doable. That said, we'd still be looking at very large pages in the project space Fritzpoll (talk) 14:01, 23 May 2008 (UTC)

Ah, I see. Well the most important thing is that the bot gets things right, so if you need to upload it to wiki, don't worry about performance. However, I'm certainly not going to be going through the 87,000-long page of China articles (:D), so only posting the lists of articles that the bot needs checked would be eminiently sensible. Happy‑melon 14:08, 23 May 2008 (UTC)

Excellent - just checking there wasn't some limit on size with article creation! It'll make for an interesting month at [{Special:Longpages]]! Fritzpoll (talk) 14:10, 23 May 2008 (UTC)

There's a maximum article length of, IIRC, 2 MB. Note that pages may have great difficulty displaying even for smaller sizes, and it will potentially be a big pain to edit (browsers can freeze up, etc.). Dumping the data onto multiple pages that are less than, say, 100 KB in size might be best. Also, Special:LongPages only counts pages from the main namespace. —Simetrical (talk • contribs) 18:23, 23 May 2008 (UTC)

Thanks for the tip-off - some of these pages may well come out to be this big. I shall split them into smaller chunks as you suggest Fritzpoll (talk) 00:16, 24 May 2008 (UTC)

[edit] Searching links to pages

Is there a way to search the pages in the project that link to a given URL, either as reference of external link, instead of the name? For example, not to search "Google" but "http://www.google.com/". Benito Sifaratti (talk) 20:11, 23 May 2008 (UTC)

Yes, you can use Special:Linksearch. mattbr 20:20, 23 May 2008 (UTC)

[edit] Link not linking

At Hicks Withers-Lancashire I have tried to insert a wikilink to Charge of the Light Brigade. However, it shews up as [[Charge of the Light Brigade]]. I have no idea why. DuncanHill (talk) 20:22, 23 May 2008 (UTC)

You had a line break stuffed in the link, breaking the syntax. --—— Gadget850 (Ed) ^talk - 20:25, 23 May 2008 (UTC)

Thanks you! DuncanHill (talk) 20:35, 23 May 2008 (UTC)

[edit] Scripts disabled until autoconfirmed?

Okay, I know that Twinkle is disabled on accounts until autoconfirmed. However, I've created this account and none of the Gadgets work either. Plus, scripts added to monobook.js won't load. Are scripts entirely disabled until the account is autoconfirmed? — The Hand That Feeds You:^Bite 19:40, 20 May 2008 (UTC)

Are you sure you have Javascript enabled in your browser? Mr.Z-man 23:46, 20 May 2008 (UTC)

Positive. This is actually a new account (I'm retiring the old one), so I've tried bypassing my cache and checking my browser settings, as well as logging on with a different browser (and on a different machine). If I log into my retired account, everything works, it's just the new one that's having trouble. -- — The Hand That Feeds You:^Bite 13:00, 21 May 2008 (UTC)

Well, I'm now autoconfirmed… and suddenly all my scripts work. So yeah, apparently scripting just plain won't work at all until autoconfirmed. Not sure if this qualifies as a bug or a feature. — The Hand That Feeds You:^Bite 17:22, 24 May 2008 (UTC)

[edit] Playing with stuff in the edit window

Is there anything I can add to my monobook.js to move the edit summary and commit/preview/changes buttons above the diff and preview when undoing an edit, but only when undoing? Also, any way to make the list of templates transcluded which is at the bottom of the edit screen collapsible (and collapsed by default)?? Happy‑melon 11:13, 21 May 2008 (UTC)

if(queryString('undo') && queryString('undoafter')) addOnloadHook(undoStyle)
function undoStyle() {
  var wp = document.getElementById('wikiPreview');
  moveObj(document.getElementById('toolbar'),wp);
  moveObj(document.getElementById('editform'),wp);
}
 
function moveObj(obj,parent) {
  var newobj = obj.cloneNode(true);
  while(obj.firstChild) obj.removeChild(obj.firstChild)
  obj.parentNode.removeChild(obj);
  parent.appendChild(newobj);
}
 
function queryString(p) {
  var re = RegExp('[&?]' + p + '=([^&]*)');
  var matches;
  if (matches = re.exec(document.location)) {
    try { 
      return decodeURI(matches[1]);
    } catch (e) {
    }
  }
  return null;
}

This relies on the query strings undo and undoafter present to detect undos, so doesn't work on the preview after an undo (also because it sticks the edit form into the preview, being the best static div available at the correct location). Giver a try --Splarka (rant) 08:01, 22 May 2008 (UTC)

Doesn't seem to do anything - is my implementation right, or do I need to wrap that code in something? Happy‑melon 09:26, 22 May 2008 (UTC)

Hmm, works for me. Cleared your cache? got any JS errors? what browser? --Splarka (rant) 10:47, 22 May 2008 (UTC)

IE7 on XP, it's not telling me it's got any errors. I know the cache is up to date because the code I added to move the search box up a section is displaying correctly. Any ideas? Happy‑melon 11:28, 23 May 2008 (UTC)

Okay, try this version dumbed down for IE7. If that doesn't work, try it at the top of your script. If that doesn't work, I dunno! I don't have IE7 handy ^_^. Remember, this only works when undoing, and only before preview. --Splarka (rant) 09:45, 24 May 2008 (UTC)

[edit] Hiding templates with a skin or something

Resolved.

As more of a reader than editor, I sometimes find templates -- especially en masse -- make that difficult. Is it possible to tweak a skin, or make any workaround, which would allow selected templates to not show up in my view? TransUtopian (talk) 02:00, 23 May 2008 (UTC)

Yes, see m:Help:User style#Non-display.--Patrick (talk) 05:43, 23 May 2008 (UTC)

Thank you, but after reading that page and m:Help:Cascading style sheets, I'm still in the dark about how to hide, for example, [citation needed] from my view. I added #fact {display: none} (and tried it with a . instead of # as well) to User:TransUtopian/Special:Mypage/monobook.css, cleared my cache, reloaded, and the template still appears.

I'm making a newbie mistake, but haven't found the aha! code that'll resolve it. More help please? TransUtopian (talk) 14:28, 23 May 2008 (UTC)

Looking at {{fact}} (and hitting my head against a wall until I understood what was going on) reveals that you want .Template-fact {display:none} In general, you can look at templates' source code to see if they wrap their contents in a named CSS class. Algebraist 14:45, 23 May 2008 (UTC)

I tried it and it's not working for me. I also tried .Template-Fact {display:none} and .noprint Template-Fact {display:none} since the scource code's class=noprint Template-Fact. I cleared the cache after each save, then reloaded a page with fact templates. TransUtopian (talk) 19:47, 23 May 2008 (UTC)

The page name should be User:TransUtopian/monobook.css (which you can also access through Special:Mypage/monobook.css).

.Template-Fact {display:none} is correct.

Patrick (talk) 22:54, 23 May 2008 (UTC)

All fixed: should work now. :) Nihiltres^{t.l} 04:25, 24 May 2008 (UTC)

Thank you very much, Patrick, Algebraist and Nihiltres! It works! TransUtopian (talk) 20:06, 24 May 2008 (UTC)

[edit] Quick monobook.css question

I'm having trouble visually distinguishing unvisited links in my watchlist. Attempting to color visited links brown, I added A:visited {color: brown;} into my monobook.css. I see no effect from this. I'm clueless; can someone please give me a clue?. -- Boracay Bill (talk) 06:04, 23 May 2008 (UTC)

Try body a:visited {color:#773333;}. Your problem might be: specificity (fixed by specifying body descendant), case (changed to lowercase a), or named colors (changed to code). Be sure to clear cache too. --Splarka (rant) 07:19, 23 May 2008 (UTC)

The issues I've always had with user subpage CSS have always been solved using the !important flag. --MZMcBride (talk) 04:20, 24 May 2008 (UTC)

I've got it working now with the change I originally made. I cleared the cache, logged out and in, restarted my browser, stood on my head while whistling "Dixie, etc. without fixing it but it eventually came good. Server update latency? I dunno. Anyhow, it's working for me now. I settled on olive (#808000)for the visited-link color. My O'Reilly HTML & XHTML book says that's a defined color name in the HTML/XHTML standards and it provides a good contrast for me in the watchlist. -- Boracay Bill (talk) 08:08, 24 May 2008 (UTC)

[edit] Table editing tool

Is there an existing tool that can reorder columns in MediaWiki tables? I'd like to move the website and MOE columns to the end of the tables in List of schools in Northland, New Zealand and the 16 other articles in List of schools in New Zealand, as suggested in Wikipedia:Featured list candidates/List of schools in Northland, New Zealand (Matthewedwards' second-to-last suggestion), but doing it manually would be rather painful. I suspect that exporting the table to a format such as xls, changing the column order, and then reimporting might be the way to go. Alternatively, I could write a python script to do the work. What's the easiest way?-gadfium 09:16, 24 May 2008 (UTC)

I wrote a python script. It only took a few minutes to write.-gadfium 22:16, 24 May 2008 (UTC)

[edit] Wikipedia:Articles written by a single editor

I do not know whether this is the correct place to post this, but Wikipedia:Articles written by a single editor comes up with a totally blank page, on two different computers, with different operating system, different browsers, and on different networks. Purging the page cache does not seem to help. — Crowsnest (talk) 14:55, 24 May 2008 (UTC)

Same for me. xenocidic ( talk ¿ review ) 14:59, 24 May 2008 (UTC)

Same for me as well. Normally when that happens, it's because it's causing a server timeout somehow, due to complicated templates that the server doesn't realise are invalid. Viewing the page with action=raw doesn't turn up any sort of complicated templaty stuff; the unusual things about the page seem to be that it's very long and contains a lot of links. Maybe that's the problem? I can imagine, for instance, that the query that calculates whether links should be redlinks or bluelinks gets mis-optimised by the database when there are a huge number of links on a page, and it does a filesort or something by mistake. This is a bug whatever's happening; maybe a dev with database access should take a look at it, if it is a database problem. Splitting the page up based on its raw text ([4]) is one possible way to solve the problem, I think; use the raw text to make subpages, then get rid of the messed-up original by deleting it or something. --ais523 15:34, 24 May 2008 (UTC)

Previous revisions of the page all have the same problem going back to the one after this one. —Ashanda (talk) 15:45, 24 May 2008 (UTC)

(ec) The talk page opens up ok... it seems from comments on there, that it is indeed the size of the page that is causing the problem. --NJJ.Rocher (talk) 15:49, 24 May 2008 (UTC)

I cut down the page size to the point where it seems to always load. When the server showed a blank page, it returned HTTP 200 if the server was running PHP 5.2.1, but returned HTTP 500 if the server was running 5.2.5. Odd. --- RockMFR 17:33, 24 May 2008 (UTC)

[edit] Counting reversions

I think a friend showed me a tool that counts number of times an editor has been reverted. I have lost it now. Does it exist? Can anyone help me? Thanks.--Filll (talk | wpc) 17:41, 24 May 2008 (UTC)

[edit] Image transparency on coloured backgrounds

Hello,

I'm an editor on the Dutch wikipedia, and it seems that our wikipedia doesn't properly show images with a transparent background (for example: ).

When this image is inserted in a table and the background is set as red (or any other colour) on this wikipedia, it correctly doesn't show any white box around the star.

The problem is, that on the Dutch wikipedia it does show the white box, even though the wikiformat is exactly the same. See nl:Wikipedia:De kroeg#Probleem met een transparante achtergrond.

I've been told it's a problem with my browser (IE) but that doesn't explain the differences between these two wiki's. I hope somebody could help find an explanation, or even better, a solution for this problem.

- Sir Iain 213.10.202.180 (talk) 18:40, 24 May 2008 (UTC)

You're using Internet Explorer 6, which does not handle PNG transparency... unless you're on the English Wikipedia. We have a script here that enables IE6 to properly handle PNG transparency. I have been badgering the Dutch admins to implement that script on the Dutch Wikipedia as well, but they don't seem interested.

You can however copy that script to your personal monobook.js. Have a look an nl:Gebruiker:Edokter/monobook.js and copy the PNGFix script. — Edokter • Talk • 19:12, 24 May 2008 (UTC)

[edit] A years revert in one go

I made this edit to revert vandalism but somehow manged to go back an entire year to this. My first thought was that I had two tabs open and picked the wrong one but, even at 500 differences a time, that would have ment going back 2 or 3 pages to get to it and I wouldn't have done that. Any ideas why this would happen? Thanks. CambridgeBayWeather Have a gorilla 21:21, 24 May 2008 (UTC)

Happens regularly. Typical scenario: you were looking at Special:Contributions/142.24.19.113 for recent vandalism, and brought up [5] without noticing it was May 2007, not May 2008. If you then click edit for the version before that vandalism, you restore everything else in that old version. Gimmetrow 21:30, 24 May 2008 (UTC)

Obvious now you point it out. Thanks. CambridgeBayWeather Have a gorilla 22:24, 24 May 2008 (UTC)

[edit] Wrapping images with text

On Myanmar, the SVG image in the lead is followed by a comma. If the page is the right width, this comma gets wrapped to the next line without the image. This is wrong; the comma and the image should be "non-breaking". Is there a solution to this? BigBlueFish (talk) 22:37, 19 May 2008 (UTC)

I added <span style="white-space:nowrap"> around it. Tra (Talk) 22:50, 19 May 2008 (UTC)

Actually you have no way of knowing how that will look in proportion to the viewer's default font size. It would be preferable to use the Unicode representation of the Burmese letters ဴပည္ေထာင္စုဴမန္မာနုိင္ငံေတာ္ rather than uploading an image for every proper non-latin name that gets mojibaked by IE. — CharlotteWebb 23:32, 19 May 2008 (UTC)

That string gets "Mojibaked" by Firefox 3.0b5 on Ubuntu, too. —Simetrical (talk • contribs) 14:47, 20 May 2008 (UTC)

That would work, other than the fact that it... um... doesn't work here either. (FF2.0.0.14/WinXP) ~user:orngjce223 how am I typing? 02:11, 26 May 2008 (UTC)

Sadly, Myanmar text gets borked pretty much everywhere because font support for it is rare, especially since its Unicode representation was recently restandardized in Unicode 5.1. The problem is compounded by Internet Explorer's refusal to try glyph substitution when the list of fonts does not include a font that supports Myanmar characters.

I'm pretty sure there is an Ubuntu package for Myanmar fonts, but I don't remember what it is. The best thing to do here is to use an SVG image to show the characters, but provide a Unicode representation in the images' alt text. —Remember the dot ^(talk) 02:51, 26 May 2008 (UTC)

I've been wondering for a while why Windows XP does not supported Burmese script, even though it was in Unicode 3.0. If you know details of what went wrong with Burmese in Unicode 3.0, please add a paragraph on the subject to Burmese script.  Randall Bart   Talk  03:35, 26 May 2008 (UTC)

I may have this wrong, but Help:Multilingual_support_(Burmese)#Encoding_models mighthave the info you're after. -- Boracay Bill (talk) 03:47, 26 May 2008 (UTC)

[edit] set focus

Hi can some one set focus on the search field. That will make my life much easier! Thanks —Preceding unsigned comment added by 65.110.174.162 (talk) 14:46, 23 May 2008 (UTC)

This is not likely to happen, as consensus has always been against it in the past. You can shift focus to the search box easily with a browser shortcut (probably alt-f, ctrl-f, shift-alt-f, or shift-esc-f), or you can register an account and do it automatically with JavaScript. See WP:FAQ/Main Page#Why doesn't the cursor appear in the search box, like with Google? for more information. Algebraist 14:53, 23 May 2008 (UTC)

If you register an account, you can easily go under your Preferences→Gadgets and select "Focus the cursor in the search bar on loading the Main Page." — The Hand That Feeds You:^Bite 15:40, 23 May 2008 (UTC)

Or hit "Tab" before typing. —Simetrical (talk • contribs) 18:23, 23 May 2008 (UTC)

That doesn't work on FireFox or IE. Algebraist 11:29, 24 May 2008 (UTC)

You're quite right. I wonder why I thought it did? It would make sense, certainly. —Simetrical (talk • contribs) 22:04, 25 May 2008 (UTC)

[edit] Wikipedia:Noindexing Talk Spaces

This proposal has been advertised widely on the Village Pump and the Community Portal, and has attracted a broad consensus of support. I would request that a developer implement this at this point. Thanks.--Pharos (talk) 20:08, 23 May 2008 (UTC)

There appear to be like half a dozen people commenting there, not a "broad consensus of support". This has been discussed in the past and it's been agreed that the value of Google for Wikipedians to search Wikipedia talk pages outweighs any mostly-hypothetical concerns about talk or project pages. A variety of pages are already blocked, including XFD, on a case-by-case basis. Google appears to block the article talk namespace of its own accord, as well.

If you really think there's enough agreement, you can file a request, product Wikimedia, keyword shell. But I don't think you've produced anything that approaches consensus, by this wiki's standards. —Simetrical (talk • contribs) 22:09, 25 May 2008 (UTC)

[edit] "You have new messages (last change)" persists for IPs

If someone leaves a message on the Talk page of an anon, the anon gets the "You have new messages..." thing. But even after having clicked "last change" link and reading the message, it often continues to be presented to the anon. (Sometimes it doesn't though - I've never quite worked out the circumstances.) I can't see the purpose in it persisting, and indeed it must be quite disconcerting to inexperienced users who naturally would start off as anon (IP) users. It's as if the record of an IP clicking on the "last change" doesn't flow through the system quickly enough. Hope this can be fixed--82.148.54.195 (talk) 13:33, 24 May 2008 (UTC)

I don't know if it's deliberate or not, but this "feature" might be useful in cases where the address is being used by several people, and the first person who sees the message notification is not the intended recipient.

I do see your point here - shared IP addresses can result in IP contributors not seeing messages to them if another user at the same IP address clicks on it first.--87.252.35.197 (talk) 19:40, 24 May 2008 (UTC)

But the larger issue is that organization of talk messages and other conveniences are geared toward logged in users. I'm a supporter of unregistered editing because I believe it encourages newbies to join us, but I think of non-registrated status as existing just for that purpose. If someone makes an edit because they stumbled upon a WP page while google searching, and just couldn't stand to see an error left uncorrected, that's fine. But if you're going to be a regular contributor, and use your talk pages, you should get an account, IMO. Making technical changes in response to requests from non-members seems to be a waste of time; no offence. --A Knight Who Says Ni (talk) 15:12, 24 May 2008 (UTC)

But I personally have to use a wide variety of computers (mostly not mine) in order to edit. I do not want to log in with my username because I do not want to leave my login details on the computers. It's usually not reasonable for me to wipe the browsing history etc.--87.252.35.197 (talk) 19:40, 24 May 2008 (UTC)

Logon details and browsing history are two different things. When on a sign-in page, most browsers will ask if you want this computer to remember this password, and if you don't, it should not be possible for someone else to find it, as long as there is no keystroke logging going on. Even if you edit from a workplace where they monitor your computer use, they should not be able to capture passwords from secure pages. If you are still afraid of your password being stolen, you could try changing it often. I'm not saying all this just to persuade you to use an account, but I do note that if you're continually switching computers, then you must also be continually switching IPs, which means you can't possibly be getting messages sent to you on IP talk pages - which is what you are concerned about. --A Knight Who Says Ni (talk) 21:29, 24 May 2008 (UTC)

If it happens again, someone should report this on bugzilla:9213. That bug's supposed to have been fixed ages ago. (9213's far from a waste of time, because it also happened in reverse, causing anons to not receive messages they should have received; that's detrimental to vandalfighting.) --ais523 15:15, 24 May 2008 (UTC)

Thanks ais523. I don't totally understand the technical details in 9213, but I'm glad it's been addressed previously. I will monitor these messages a bit more closely in future and raise this again if it recurs. (Funny things happen from time to time if you mostly edit as an anon like me.)--87.252.35.197 (talk) 19:48, 24 May 2008 (UTC)

To clarify, quoting Brion in comment 33 of that bug:

new messages notification is only shown if a session cookie is present, indicating that a private session is in progress and pages won't be shared. This generally means that you've either hit an 'edit' link or the login button at some point in your browsing session.

If you have no cookies, you will not receive a "new messages" bar, and this will probably never be changed, for performance reasons (but you'll get the message as soon as you try to edit). However, people shouldn't be getting new messages bars persistently, AFAIK. If you could get the full HTTP headers from the response, using a tool like Firebug (Tools -> Firebug -> Open Firebug -> Net -> click on first request -> Headers -> copy and paste the "Response Headers" section), that might be helpful to someone. —Simetrical (talk • contribs) 22:16, 25 May 2008 (UTC)

[edit] unwanted slider

see Richard Durbin, an article I was looking at. On the right side of most of the article is an up/down slider which overlays the right-most characters of the article. I do not see what is causing this/how to fix it. Can anyone fix it? Hmains (talk) 17:50, 25 May 2008 (UTC)

Could you tell us what browser and which browser version you are using? --TheDJ (talk • contribs) 19:04, 25 May 2008 (UTC)

Some kludgy divs to make the table horizontally scroll. I fixed it. — Edokter • Talk • 20:41, 25 May 2008 (UTC)

[edit] Why are some items on watchlist showing as bold and items in history have odd message?

Some items have just recently started appearing on my watchlist as bold. I've no idea why. Also, when I look at the edit history for page that I was the last person to edit, the following message, highlighted in light green, appears after the edit summary for my edit: updated since my last visit. Again, I've no idea what to make of this. older ≠ wiser 02:08, 19 May 2008 (UTC)

It means they were updated since your last visit. However as it's a bit flaky atm, it's been disabled again for now. --brion (talk) 02:09, 19 May 2008 (UTC)

Still showing up as bold for me. I've also tried visiting a page and coming back to the watchlist, but it seems as if it's not unbolding it correctly. BuddingJournalist 02:11, 19 May 2008 (UTC)

Also still present for me. If this is gonna be a longterm change, can you tell me the necessary style information now so I can block it in my monobook (it makes the Watchlist HIDEOUS)? JPG-GR (talk) 02:14, 19 May 2008 (UTC)

It might be a good improvement once it works okay, but at the moment it is fairly confusing. I think the disabling may have bolded everything rather than unbolding everything... Geometry guy 02:14, 19 May 2008 (UTC)

(after ec) Thanks. The bolding on watchlist, perhaps I can see that. But the message in edit history was on my edit. There was no subsequent edit (even after purging). older ≠ wiser 02:12, 19 May 2008 (UTC)

Yeah, the bolding of titles, I'm not exactly sure where the helpfulness is, I check the time since I last clicked on my watchlist and go from there. The green "updated since", I saw on commons and spent awhile trying to figure out, since no-one had done anything to the page, and I couldn't even find a linked page having been deleted. Confusing at best, can we have a preference here? Franamax (talk) 02:16, 19 May 2008 (UTC)

I'll second that. If this is permanent, I'd like to have the ability to individually disable it. -MBK₀₀₄ 02:17, 19 May 2008 (UTC)

This is a great improvment. Watchlists on other projects have this feature (ie Commons, meta, books) and I was wondering when Wikipedia would catch up to have this useful feature, I hope it stays. -- penubag  (talk) 02:18, 19 May 2008 (UTC)

Hmmm no wonder the logo hasn't been changed yet. I really wish the community-proposed developments with consensus were given more priority than these backroom things they do. It's really quite frustrating. Equazcion •✗/C • 02:19, 19 May 2008 (UTC)

Is this going to be fixed any time soon? Everything is in bold for me as well. My left eye has melted out of my skull. Gwynand | Talk•Contribs 02:21, 19 May 2008 (UTC)

Cool. Got a pic of that? Raymond Arritt (talk) 02:24, 19 May 2008 (UTC)

Some code changes cleanup improper settings checks. The downside is that it triggered some old code that bolded items by default, since apparently, that was how they were in 1.4. Items should be in regular font now. Aaron Schulz 02:27, 19 May 2008 (UTC)

To summarize:

1. The bold-since-last-change is now disabled here again as it doesn't seem to work quite properly. (It used to be inactive always unless e-mail notification options were enabled.)
2. Disabling it temporarily caused the behavior that *everything* would be bold. This was apparently a weird "compatibility" mode for a 4-year old version of the software, and has been fixed. :)
3. If and when the markers work consistently, they'll be enabled. You won't have the option to opt out of them as a preference, as it should be very clean and transparent, but of course it would be trivial to do a CSS hack to remove the bolding if you can't stand having the information available. :)

--brion (talk) 02:27, 19 May 2008 (UTC)

Well, I'd recommend having that CSS hack ready... :) JPG-GR (talk) 02:33, 19 May 2008 (UTC)

I appreciate this whole boldness in applying boldness thing, but lots of people hate it, and there's near-unanimous community consensus that the logo should be changed. Can't we do the stuff people actually want and have discussed first? I hate to complain and stuff, just saying. This kinda came out of left field. Equazcion •✗/C • 02:36, 19 May 2008 (UTC)

Got a new version ready? If so, file a bug request with: a link pointing to the updated file, a link pointing to the on-wiki consensus, and use the 'shell' keyword. You'll have far better odds of having something accomplished that way than kvetching here. ; - ) --MZMcBride (talk) 02:53, 19 May 2008 (UTC)

It's all done already, a few days ago: bugzilla:14137. More generally though, requested things should be given priority. I mean we;ve got the edit conflict bug on file for years, bugzilla:4745, and they're working on this bold crap? I'm sorry this is just really ridiculous. Equazcion •✗/C • 02:56, 19 May 2008 (UTC)

The majority of the developers are volunteers and work on what they feel like/have experience in/think they can do. I believe Aaron has been working on this for at least a couple weeks now (or at least he was talking about it), far before the proposal to change the logo began. Further, only a handful of the developers have shell access to the servers and actually can change the logo. Mr.Z-man 03:18, 19 May 2008 (UTC)

Understood, but I'm a volunteer too. Developers are also part of the community, aren't they? If someone can point me to a community discussion that addressed the need for bold links in the watchlist with enough support was gained, I'll shut up. But as a frequenter of VP proposals, I find it frustrating that we can discuss things til we're blue in the face and nothing that requires developer intervention gets done, while they don't discuss anything with us and just implement what they like. Equazcion •✗/C • 03:23, 19 May 2008 (UTC)

A) Something that's easy, like the logo will get done soon, its not like its urgent. More difficult things like bug 4745 will wait until someone actually comes up with a solution, just because there's consensus that it should be done does not mean that it necessarily can be done. B) The English Wikipedia is not the only project. The majority of software changes (like the watchlist) will affect more than just us. C) If every project had to get consensus for every software change, we wouldn't have much. Remember rollback? Mr.Z-man 03:33, 19 May 2008 (UTC)

I'm not asking for wide consensus, but some discussion to determine whether this is worth the time or even whether it's a popular feature would've been nice. With everything we're currently waiting on, it's frustrating to see something being implemented that wasn't even discussed. It seems like the software is someone else's pet project and the community isn't a priority. If the developers have a good feature suggestion, why not bring it up at VP to see what we think before they spend 2 weeks developing it? Equazcion •✗/C • 03:41, 19 May 2008 (UTC)

It's quite difficult to instruct the volunteers. One can always step in to do it yourself. Beyond that, bitchin' and badgerin' aren't productive strategies. Has anyone said yet how happy we are that the dev's have produced this new feature? Way to go people! Some of us might think it sucks, but we probably haven't had a good look yet, and thanks for your help :) Now, having done that great piece of work, here's another one to look at... Franamax (talk) 05:31, 19 May 2008 (UTC)

Yep, Meta, Commons, and the rest of the people who asked that Enotif be merged into the trunk a couple of years ago. Titoxd^{(?!? - cool stuff)} 03:53, 20 May 2008 (UTC)

Please note, first of all, that most developers do not have shell access and could not change the logo. Aaron, who I take it was doing the fiddling that caused what some people are complaining about, would not have been able to do what you ask even if he wanted to. Second of all, developers who are volunteers (i.e., all but a couple) will do what they feel like working on. Saying that they have a duty to work on what the community wants is like saying all Wikipedians have a duty to work on the articles selected by WP:AID. They don't (although we might have more featured articles if they did!). The one or two who are paid, of course, are not directly responsible to the community either, but rather to the CTO and ultimately the Board, like any other paid Wikimedia employees.

I should also point out that shell work like changing logos is extremely boring and repetitive (keep in mind that there are hundreds of wikis that make requests like this on a continuous basis; see the 112 open bugs of this sort) and only a very small number of people are able to do it. In the present setup, anyone able to perform this kind of change has root database access. This gives them the rights of checkuser, oversight, steward, except without logging (look here and see if you can find any rights logs). And they get abilities far beyond those, too. They can silently and untraceably alter any aspect of any article's history, logs, or whatever else they want. They could in fact delete the site, out of either malice or incompetence. Now to compound all this difficulty, enwiki has the unique habit of going ballistic when a request is fulfilled and they then decide that they really hadn't made up their mind after all (witness: rollback), so some shell users are probably reluctant to do what it asks for.

Anyway, back to the question of who gets to decide what. The goal of developers is not to do anything for the English Wikipedia. It's to improve the software, for all users. It is not remotely practical for developers to start a discussion on every single Wikimedia wiki, especially since most are non-English. German and French Wikipedia editors have their opinions just as much as English Wikipedia editors do; what purpose would an enwiki discussion alone serve? We could have the discussion on Meta, but that still misses the point. MediaWiki is one of the most popular wiki software packages there is. It's used by hundreds (thousands?) of Wikia wikis, and countless others scattered around the Internet. We can't go out of our way to ask all of them their opinion before we make any change. Nothing would ever get done.

Now, we could still have discussions somewhere before implementing changes, and in fact we do. Discussions about software features regularly occur on wikitech-l and #mediawiki. If you're interested in keeping track of this kind of thing, you need to go there, not expect us to treat enwiki specially and come to you before doing anything. In many cases changes are still committed without any discussion, but those who follow mediawiki-cvs-l can comment on them before they go live. If you choose not to do so, then don't complain when you can only object after they go live. You can't reasonably expect MediaWiki development to give any special preference to the English Wikipedia. —Simetrical (talk • contribs) 14:41, 19 May 2008 (UTC)

• (ec, this was in reply to franamax) That's just it, I'm not too happy that this was done. The fact that a lot of work went into it doesn't say anything about whether or not it should've been done. If I spent a month writing a policy in my userspace that would eventually affect everyone on Wikipedia, and I suddenly moved it to WP: space and tagged it as a policy, everyone would tell me listen, we love you for putting in all that work, but you should've discussed it with us first. I'm telling you (everyone) the same thing. I appreciate all the work you put it, but sorry as I am to say this, it should've been discussed first. You're part of the community, or at least you should be, and you're here for them the same way I'm here for them. The processes for implementing changes that affect everyone apply to you too. Just because you have the ability to circumvent that is no reason to do it. Equazcion •✗/C • 14:46, 19 May 2008 (UTC)

• In reply to Simetrical: If this is the way it's supposed to work, and the community isn't actually supposed to be involved with development, with its discussions occurring away from them, then I'll have to accept that. I don't agree with it, but if this instance is just one example of the way things have always worked, then my beef is with the way things have always worked, and I apologize for complaining about this one thing.

• In response, to "developers who are volunteers (i.e., all but a couple) will do what they feel like working on. Saying that they have a duty to work on what the community wants is like saying all Wikipedians have a duty to work on the articles selected by WP:AID" -- This is different from telling me I should be working on a requested article rather than whatever article I want. We're not talking about articles, but changes that affect everyone. If there were a pressing and demonstrated need for a particular policy, I'd feel obligated to offer my assistance there. Similarly I think requests from the community (yes even from one language version) should be given more priority than a feature that was just discussed among the developers. Again, developers are here for the community, the same way I'm here for the community. I think they're treated (and treat themselves) as too much of a separate entity. I appreciate all the work they put in, but if the end goal is to help write the encyclopedia, and the people doing the writing have already decided on something that would help them do that, it should be given priority over other things. Equazcion •✗/C • 15:25, 19 May 2008 (UTC)

  You might feel obliged to offer your assistance if there were a pressing and demonstrated need for a particular policy. Likewise a developer might feel obliged (and often they do) to work on something that's viewed as important. But in both cases, it's the choice of the volunteer, not a duty. And in both cases, it's the volunteer who decides whether it's pressing and needed, not the community.

  Anyway, the end goal of development is not just to build the encyclopedia. That's precisely my point. You're being too Wikipedia-centric, viewing MediaWiki as just an extension of Wikipedia. It's not. MediaWiki is a piece of software in its own right, and the goal of the MediaWiki project is to improve it for all users, present and future. That includes, most notably, the English Wikipedia, but that has no special status at all, except as far as individual developers may or may not care about it more.

  Developers are not here for the Wikipedia community, they're here to develop MediaWiki, for whoever they want to prioritize. We have at least a couple of committers from Wikia, who theoretically work for Wikia. More than a couple of developers run their own wikis and do a considerable amount of coding to benefit those. Many features that are added will never be turned on on Wikipedia, and the author knows that. Many developers write extensions and features that are solely of benefit to non-Wikimedia users. As it happens, a large percentage of developers are particularly interested in the English Wikipedia, and may choose to spend their time working on stuff for it. But the end goal is not just to help write the encyclopedia. —Simetrical (talk • contribs) 14:44, 20 May 2008 (UTC)

Please keep in mind that more time and effort has been put into this thread than the recent one-line tweaks to the bolded-pages-in-watchlist feature (which has been in the software for years, but originally was attached to the e-mail notification feature which isn't enabled here yet in part for performance reasons). --brion (talk) 16:04, 19 May 2008 (UTC)

Hey, I'm just a volunteer. Equazcion •✗/C • 16:08, 19 May 2008 (UTC)

 :D --brion (talk) 16:12, 19 May 2008 (UTC)

Yes, but bickering is much more fun than coding. —Simetrical (talk • contribs) 14:44, 20 May 2008 (UTC)

Not all discussions are bickering, and if coding myself were an option, I'd be doing that rather than trying to push others to. I'll also note that about half the content of this exchange consists of your responses, Simetrical. Equazcion •✗/C • 18:16, 20 May 2008 (UTC)

You should be glad I am not well-versed in such matters... This is nothing. :-D Waltham, The Duke of 01:09, 21 May 2008 (UTC)

I was remarking on my own tendency (which I'm sure Brion has noticed before) to talk more than code, not disparaging you. But as for you, you always have the option of learning how to code if you're willing to put in the time and effort. —Simetrical (talk • contribs) 14:07, 21 May 2008 (UTC)

Oh, no, not what I meant at all. I just made a remark on my verbosity; had I been able to participate in the discussion, it would be twice as long.

As far as the time and effort are concerned, this is a relative concept. I need yet another thing, though: resources. Where could I go to find material to study? Waltham, The Duke of 00:43, 27 May 2008 (UTC)

Wikiversity? And I'm sure there are lots of computer scientists floating around Wikipedia that you could ask for more specific help. If you've never done programming before then it would be a good idea to start simple, with Visual Basic .NET, Python, or C++. —Remember the dot ^(talk) 01:08, 27 May 2008 (UTC)

[edit] See a user's contributions to a specific page?

I know how to see a user's contributions to a specific namespace, but how can I see a user's contributions to a specific page? Or is the best way to do this is to show a user's contributions to a namespace, show as many as I can on one page, and then do a normal Find command to find the page I'm looking for? Gary King (talk) 14:16, 26 May 2008 (UTC)

See #Search For User Contributions On A Specific Talk Page, Article Page, Category Page, Etc. Etc. and Wikipedia:Village pump (proposals)/Persistent proposals#History Options for previous discussion. Algebraist 14:20, 26 May 2008 (UTC)

Hm, the API can be used to do this, per one of the replies... Gary King (talk) 14:22, 26 May 2008 (UTC)

Yes, you can use the API to get a list like this. See for instance this example, which should list all MY edits to the page STS-118 --TheDJ (talk • contribs) 15:01, 26 May 2008 (UTC)

[edit] Performance?

Resolved. Works fine now, whatever it was is gone away. Franamax (talk) 21:50, 27 May 2008 (UTC)

Is it just me or is the whole site falling apart last hour or two? Some of my edits sail through, some hang up forever. I've just installed XP SP3, so I'm a little concerned. Thx! Franamax (talk) 21:22, 24 May 2008 (UTC)

I've had the server error a couple of times today, but I wouldn't say the site was significantly slower than usual. Happy‑melon 22:02, 24 May 2008 (UTC)

What browser are you using? —Remember the dot ^(talk) 22:11, 24 May 2008 (UTC)

IE7 7.0.5730.11 (latest update). Either it's the SP3, or Brion forgot to schedule extra kitties for the weekend. Franamax (talk) 22:23, 24 May 2008 (UTC)

That last edit took 52 seconds to return, and another 10 seconds to complete the page update. That's not normal... Franamax (talk) 22:27, 24 May 2008 (UTC)

[edit] Sortable wikitable: anomaly when sorting on numbers - minor bug?

Resolved. Gary King (talk) 15:40, 27 May 2008 (UTC)

On Eurovision Song Contest 2008, the "final" table (I mean the one listing the final positions with Russia first, UK last) sorts correctly the first time you click on the Place column's sort button. But when you click on it several times, it then seems to do an ASCII sort when ascending, ignoring the fact that they are numbers. I think this is because there are asterisks against some of the numbers. Strangely, it always sorts as expected when descending. Is this a bug? Or should we never try to put asterisks etc. against numbers?--86.145.248.219 (talk) 08:25, 25 May 2008 (UTC)

Yes, it is due to the asterisks, use Template:Nts.--Patrick (talk) 09:23, 25 May 2008 (UTC)

Yes, I thought it was the asterisks. I can't seem to get it sorting correctly with Template:Nts, but it's probably not worth the trouble anyway.--86.145.248.219 (talk) 09:58, 25 May 2008 (UTC)

What is happening exactly is the following. On the first time you click Place, the first row contains "20". This is recognized therefore as a "number" column that is sorted according to the "number" sortmethod. Now it is ordered from 1 trough 25 in ascending order. The first row contains "1". You hit the sort button again. The first row is "1", so again number sorting is used but now in descending order from 25 trough 1. Now however the first row contains "25*". When you hit the sortbutton to go back to ascending order, due to the *, string sorting will be used. After this stringsorting "1" will be the top row, which again is a number and therefore it will now alternate between number sorting and string sorting. Using {{nts}} on the numbers with * will fix this issue. --TheDJ (talk • contribs) 10:20, 25 May 2008 (UTC)

Thanks for the explanation - the behaviour makes sense now. If someone knows how to apply {{nts}} in this case, and is willing to do so, I'd be grateful.--86.145.248.219 (talk) 11:25, 25 May 2008 (UTC)

Y Done --TheDJ (talk • contribs) 12:17, 25 May 2008 (UTC)

Thanks.--86.145.248.219 (talk) 12:53, 25 May 2008 (UTC)

[edit] WhatLinksHere loop

While disambiguating [[Nativity]], I ran into an odd problem on the WhatLinksHere page. I found myself at this URL and when I click "(previous 50)", I come back to the same page. Using the 20, 50, 100, 250, 500 links did not resolve the issue. The only escape was a link at the top of the page added by my WhatLinksHere script. I don't see how my script could be the cause of the problem, since it doesn't modify those links.  Randall Bart   Talk  18:21, 25 May 2008 (UTC)

I had a similar problem discussed at Wikipedia:Village pump (technical)/Archive 38#What links here problem. DuncanHill (talk) 18:36, 25 May 2008 (UTC)

How did you end up at that link in the first place, if using the normal link (without specific reversion bounds) it seems to be working ok. — xaosflux ^Talk 11:38, 26 May 2008 (UTC)

I can replicate a problem similar to this one by going to [6], then clicking next 50, next 50, previous 50, and previous 50. --- RockMFR 04:22, 26 May 2008 (UTC)

Ah, ok looks like there is a bug in that the "previous" page link is not encoding the right bounds in to the URL, sounds like a bugtraq ticket may be needed here. — xaosflux ^Talk 11:38, 26 May 2008 (UTC)

WhatLinksHere uses an awful paging model and should really be updated to use the Pager class we have for this. —Simetrical (talk • contribs) 15:00, 26 May 2008 (UTC)

Which Aaron has now done, should be live sometime soon. —Simetrical (talk • contribs) 15:50, 27 May 2008 (UTC)

[edit] Whitespace in table

Resolved. Gary King (talk) 15:39, 27 May 2008 (UTC)

A cookie to anyone who can make the whitespace (coffeespace? :D) between the "importance" row and the collapsible table at the bottom of the banner on this page disappear... Happy‑melon 18:46, 25 May 2008 (UTC)

Y Done by RockMFR --TheDJ (talk • contribs) 19:08, 25 May 2008 (UTC)

[edit] Login problem

This isn't a question, more of an observation that might be an indication of a problem; let me know if you think it is.

I've been using my WP account for a couple of months from my computer at home, and never have to sign in, because I have "remeber me on this computer" ticked off on my preferences. Just now, I found I am no longer logged in automatically. When i got signed in, the box is still ticked off, and my browser seems to be enabling cookies with no problem. I'm wondering if my IP address got changed; I understand that can happen every now and then with a statis address, if it expires. I don't have my old address to compare. I thought I did, but I'm using a router box, and it turns out the IP address I wrote down (via "ipconfig" from a dos window) is just for the router, and is not the address that the internet knows me by.

Anyway... when I went to log in, I noticed the login page gives an option to sign in via a secure server. I thought great, I'll use that. But after doing so, I found I have to click on 2 warning boxes that I'm going into and out of a protected page with EVERY click on a link within WP. Also, when I closed my browser and came back, I found myself logged out again. I've used secure sign-in pages on other sites, and have never seen this behavior. The secure page status should just be for the sign-in page itself, right? And regardless of which sign-in page I used, I should have gone back to automatic sign-in status.

After a couple of go-rounds, I finally gave in and used the unsecure server to sign in. I haven't checked to see if I'm still logged in after closing the browser. But I'm wondering if the every-page-is-secure thing is a known WP problem (or is it a problem at all, maybe it's supposed to do that), or if there was a system-wide hiccup today that is forcing everyone to log in, or anything like that. --A Knight Who Says Ni (talk) 22:55, 25 May 2008 (UTC)

First of all, note that the saved login cookie expires after 30 days, at which point you must log in a second time. Once a month isn't too bad. :)

As for the secure login; currently images (and possibly some local Wikipedia JavaScript) are still pulled from an unsecured connection. This may cause some browsers to show you warnings, depending on your security settings. --brion (talk) 23:39, 25 May 2008 (UTC)

Most likely candidate culprits in this case are geonotice and wikiminiatlas, I think. --TheDJ (talk • contribs) 00:08, 26 May 2008 (UTC)

When browsing the secure server on Firefox, it won't bother you with pop-ups. Instead, Firefox just shows the page to you as though it were unencrypted. It simply removes all visual cues that you are visiting a secure site because some of the content is sent unencrypted, which is far less annoying than Internet Explorer's constant pop-ups. —Remember the dot ^(talk) 01:27, 26 May 2008 (UTC)

If only it did that for bad certs, too . . . —Simetrical (talk • contribs) 15:01, 26 May 2008 (UTC)

Bad certificates are worse because they indicate a potential attempt to substitute a malicious web page for a real one. That's why Firefox (and all other browsers, I think) make sure to warn you about them. —Remember the dot ^(talk) 16:21, 26 May 2008 (UTC)

Bad certificates indicate that the web admin wasn't careful in configuring the site. I've honestly never heard of an actual malicious attack through this route (although I suppose they must have occurred) ― whereas cert mismatches due to admin error are more or less a daily occurrence, with millions of examples (like, oh, https://google.com/). Paranoid security messages simply teach people to ignore security messages. They're stupid and destructive. —Simetrical (talk • contribs) 15:54, 27 May 2008 (UTC)

Fixed; Wikiminiatlas JS now loaded over SSL. Firefox 3.0rc1 shows the lock icon & blue bar again. (Old JS may remain cached, so force reload to confirm.) --brion (talk) 17:56, 26 May 2008 (UTC)

Hackers everywhere rejoice that people ignore the bad cert warning. _{1 != 2} 16:31, 26 May 2008 (UTC)

If you go to https://cia.gov/library/publications/the-world-factbook/ or https://odci.gov/library/publications/the-world-factbook/ or https://www.odci.gov/library/publications/the-world-factbook/ you will get a cert warning, because the domain name in the cert is www.cia.gov. A cert should include all the applicable domain names, but doesn't. I suggested this be fixed ten years ago, but AFAIK the standard still allows just one. The cert issuers don't want to fix this, because they make more money when sites buy multiple certs. As long as this remains broken, cert warnings will be commonplace and they will be ignored.

Expired certs are also a problem. At many sites, the only way the admin knows that the cert is expiring is when users start complaining that the cert has expired. There may be a mechanism for warning the cert owner ahead of time, but I'm sure at most sites it goes to an email address that gets ignored.  Randall Bart   Talk  20:59, 27 May 2008 (UTC)

Last I checked, most modern browsers support the SubjectAltName extension. No idea how much the commercial CAs would charge to sign such a thing (if they even would), but it seems to work pretty well. It will even allow name-based virtual hosting in Apache, although Apache does complain on startup. Anomie⚔ 23:27, 27 May 2008 (UTC)

[edit] Language clean-up category splitting by language?

(copied from the main Village Pump page)

The categories which I try to do some work in, as a German-English bilingual, such as Category:Rough translations and Category:Wikipedia articles needing cleanup after translation, are generated by templates which have a parameter for sources language - so would it be possible to split the cleanup categories by language? It's really annoying to have to click through almost every article trying to find the ones that are from German - and having a distinct category for, say, Japanese, might make it easier to recruit Japanese-English bilingual help. Cricketgirl (talk) 08:05, 23 May 2008 (UTC)

Not a bad idea, that; and I doubt very controversial. You might want to ask over at WP:VPT to get a template coder to modify the templates and maybe a bot to sort the articles. —Ashanda (talk) 12:33, 24 May 2008 (UTC)

I can do that, but it would require some modification to the way the language parameter is passed to the templates. Currently the |language= parameter is freeform: anything, including wikilinks, is acceptable. To implement this properly, input here would have to be restricted, probably to the ISO 3166 country codes, or these plus the full name of the language. If these templates are associated with bot scripts or WikiProjects/collaborations, then more discussion is needed with them; otherwise I could do this immediately. Happy‑melon 10:52, 27 May 2008 (UTC)

I've had a quick look around - I found about of discussion that the category was full at Wikipedia talk:Pages needing translation into English, which is linked to Category:Wikipedia articles needing translation, but neither of the other categories I mentioned has discussion pages, and I went through the WikiProject Directory and only found Pages needing translation and Wikipedia:Translation. So I think the templates are not part of a project or anything. So if you wouldn't mind, is it fairly easy to split the categories? I realise that some of the time quite a few will be empty or underpopulated, but I think it's worth it to make translation and cleanup simpler (I notice that quite a few of the cleanup categories have a "this category may be empty but it's still useful" sign - would it be possible to add that to the new split categories?) Thanks! Cricketgirl (talk) 19:01, 27 May 2008 (UTC)

[edit] New color for redlinks?

Have redlinks just been changed to a new, brighter (blood-red) color? Badagnani (talk) 07:31, 27 May 2008 (UTC)

It looks the same to me. FYI, the shade of the red changes after you click on one. Gary King (talk) 15:07, 27 May 2008 (UTC)

[edit] Parag break not working after {{cn}} tag

Anyone know why I had to make this edit to fix the glitch? --Dweller (talk) 10:16, 27 May 2008 (UTC)

No, but I had to do the same in another article last week. I reported this at Template talk:Fix. Oddly, if you copy the affected text to a sandbox, it works. A space will fix it as well as a break. --—— Gadget850 (Ed) ^talk - 14:07, 27 May 2008 (UTC)

I've tracked down the bug, it seems that the category link appended by {{fact}} (on article pages only, which is why copying it to Wikipedia:Sandbox "fixes" it) in combination with a wikilink/image at the start of the next line causes the parser to lose the break. A minimal test case:

I don't know if this is a recent change in the parser, a recent change in {{fact}}, or if it has just always been this way. Anomie⚔ 17:29, 27 May 2008 (UTC)

[edit] dn tag misbehaviour again

{{dn}} disrupts line-breaks, see [7] for an example. I recall raising this problem before, but don't recall what the outcome was! DuncanHill (talk) 10:42, 27 May 2008 (UTC)

Oh, here it is - Wikipedia:Village pump (technical)/Archive 37#Line of text shooting off the right-hand edge of the screen. DuncanHill (talk) 10:44, 27 May 2008 (UTC)

Still an open issue at Template talk:Fix. --—— Gadget850 (Ed) ^talk - 13:51, 27 May 2008 (UTC)

[edit] Autoconfirmed promotions

Is there any way to access a log or real-time feed of protions to the 'autoconfirmed' usergroup? For that matter, is there any way of directly checking whether a particular user is autoconfirmed? Happy‑melon 13:44, 27 May 2008 (UTC)

You could find out when a user registered their account then count forwards by four days? Such as, for yourself, it's here. Gary King (talk) 15:05, 27 May 2008 (UTC)

You'd have to check for ten edits as well, but yes, I am aware that it could be done like that. The reason I asked is that I want to do it in reverse: from the list of autoconfirmed users, check somethign else, not from the list of all users, check if autoconfirmed. I'm basically trying to write a script to ferret out autoconfirmed sleeper socks and vandal accounts. Happy‑melon 15:27, 27 May 2008 (UTC)

autoconfirmed is not like most usergroups as far as I know, especially since this returns an "Unrecognized value for parameter" error. Quoting from the source code itself, it is an "implicit group", similar to the group of users who have confirmed their email addresses, which belong in emailconfirmed. Gary King (talk) 15:33, 27 May 2008 (UTC)

Autoconfirmed is not stored anywhere, it's checked as required. So there are no actual promotion events to log; you'll just be checked every time for age and edit count, and nothing records whether the last check gave a different answer. —Simetrical (talk • contribs) 15:57, 27 May 2008 (UTC)

Yeah, that's what I imagined was the case after looking through the code. There are several cases where this behavior is preferred; for instance, a new user makes several edits but they are all deleted, then they end up back with 0 edits with the current way the system works. Gary King (talk) 16:00, 27 May 2008 (UTC)

The edit count doesn't go down when your edits are deleted, so that's not actually the case. --brion (talk) 16:48, 27 May 2008 (UTC)

Ah yes, that's right. Gary King (talk) 17:21, 27 May 2008 (UTC)

[edit] HTTP500 error

In the last hour, I've suddenly had editing problems: half the time when I save a page, it gives me a page entitled "HTTP500 Internal Server Error", and the page displays the following text:

The website cannot display the page HTTP 500 Most likely causes: The website is under maintenance. The website has a programming error. What you can try: Refresh the page. Go back to the previous page. More information This error (HTTP 500 Internal Server Error) means that the website you are visiting had a server problem which prevented the webpage from displaying. For more information about HTTP errors, see Help.

I used to get this message occasionally, but I'd find that the edit was saved; however, today I'm finding that no edit that results in this message gets saved. Some pages don't have any problems; on the other hand, when trying to make a single edit to Kettle River, Minnesota, I got the error three straight times before succeeding on the fourth one. Similarly, when I saved this new thread, it was the fourth time, for the first three times resulted in errors. Any idea what's happening? Nyttend (talk) 15:33, 27 May 2008 (UTC)

The server is crapping out. It's worse today than usual for me, too. Gary King (talk) 15:37, 27 May 2008 (UTC)

The error page you report is generated by Internet Explorer and has no relationship to reality. If you could give the output you get from a browser that actually prints out the HTML error page returned by the web server (in other words, use any browser but IE), that would be considerably more helpful.

Someone just reported in #wikimedia-tech that they were getting database errors, anyway, and gave the specific error (from, one presumes, Firefox or some other non-retarded browser). Domas said he fixed it, so maybe the error will go away now. —Simetrical (talk • contribs) 16:00, 27 May 2008 (UTC)

It has certainly improved for me during the past few edits. Gary King (talk) 16:01, 27 May 2008 (UTC)

[edit] Wikiwandering: logged in everywhere?

I've just been wiki-wandering and ended up at the Turkish wikipedia, which I've never visited before in my life. I noticed that I'm already logged in with my global account on that wiki, even though I've never visited tr:Special:UserLogin. Is this a new feature of SUL? If so, congratulations and thanks to whoever implemented it. Happy‑melon 16:18, 27 May 2008 (UTC)

I recall reading about it a few days ago, but yes, it is a new feature. Gary King (talk) 16:25, 27 May 2008 (UTC)

It's been active for several weeks on the SSL interface; we've just enabled it sitewide. --brion (talk) 17:45, 27 May 2008 (UTC)

Well thanks! It's a great feature. Some discussion on AN and analysis of the account creation log at fur:wiki:Special:Log seems to be suggesting that it's not working for everyone... Is the sudden influx of account merge activity the reason the whole site's running so slowly? Happy‑melon 18:42, 27 May 2008 (UTC)

[edit] Changes to log out page

There seems to be a change to the log out page. I get:

"You are now logged out. This computer may be used to browse and edit Wikipedia without a username, or for another user to log in. Note that some pages may continue to be displayed as if you were still logged in; this can be fixed by clearing your browser cache.

Return to {{Pagenamehere}}."

In the big gap, I get a lot of red crosses, like image files that aren't formatted. Anyone else getting this? D.M.N. (talk) 17:49, 27 May 2008 (UTC)

Do you have an ad blocker? -- Tim Starling (talk) 18:04, 27 May 2008 (UTC)

I have Norton Anti Virus. Besides, I didn't think Wikipedia had adverts. Shall I upload the image I get when I log out so everyone can see, and in case a lot of people are getting this? Thanks, D.M.N. (talk) 18:41, 27 May 2008 (UTC)

These are used to log you in/out globally on all our domains, if your account has been unified. --brion (talk) 20:02, 27 May 2008 (UTC)

[edit] Whatlinkshere broken, fixed

Yes, it was broken for a few minutes. --brion (talk) 20:15, 27 May 2008 (UTC)

[edit] Infobox dab problem

Template:Infobox U.S. congressional district, when Georgia is entered as the "state" value generates a link to Georgia, which is a disambiguation page. It should link to Georgia (U.S. state). If Georgia (U.S. state) is entered in the state field, the correct link is generated, but it displays as Georgia (U.S. state) and cannot be piped to display as Georgia. I think the template was changed recently causing this, but do not understand how it works and am reluctant to fiddle with it. Any help much appreciated. DuncanHill (talk) 21:50, 27 May 2008 (UTC)

Fixed. Someone removed the automatic disambiguation for the state parameter. — Edokter • Talk • 22:23, 27 May 2008 (UTC)

Thanks! DuncanHill (talk) 22:26, 27 May 2008 (UTC)

[edit] Move the search box directly beneath the puzzle globe

Currently the search box, the first thing that our users want to use, is about 3/4 of the way down the screen, after the lists of "navigation" and "interaction" links. This can make it hard to find for new and elderly users.

I propose that we move the search box directly beneath the puzzle globe, like so:

Comments are welcome in the discussion at Wikipedia:Village pump (proposals)#Move the search box directly beneath the puzzle globe. —Remember the dot ^(talk)

I actually think it's too high, because if you find yourself scrolling through an article, and do not find a link to take you where you want to go, you can get lost trying to find the box again, that is, if your browser isn't having too much demanded from it to allow you to do so. Since typing defaults as an entry into the search box, I don't see how someone just starting out on a page has nearly the same problem. All they have to finish with is enter. MMetro (talk) 21:35, 28 May 2008 (UTC)

[edit] Limiting HTML table sizes

WARNING: These links may crash your web browser and display Goatse.

• http://en.wikipedia.org/w/index.php?title=Template:Pirates&diff=prev&oldid=214242130
• http://en.wikipedia.org/w/index.php?title=Template:World%27s_most_populated_urban_areas&diff=prev&oldid=213966317

Would it be feasible for MediaWiki block attempts at inserting excessively large raw HTML tables into pages? I'm noticing a few vandals here and there who use this trick to "upload" shock images despite being unregistered and to also crash the web browsers of those who patrol CAT:RFU. --  Netsnipe  ►  20:24, 22 May 2008 (UTC)

According to Template:Pirates' history page, the article size was 1,975,193 bytes after the vandalism. Template:World's most populated urban areas was 1,077,378 bytes. Perhaps a simpler solution is to restrict to the total article size to 500KB, or a similar size... Rami R 14:40, 23 May 2008 (UTC)

That wouldn't stop all of it, but at least it would stop the largest pages, whether goatse, or repeated text. Page sizes are limited on some wikis (not sure which) but I don't know if they use the same software as Wikipedia. --Snigbrook ^(talk) 01:44, 24 May 2008 (UTC)

And large page sizes are a problem even if the content is useful. --Snigbrook ^(talk) 01:46, 24 May 2008 (UTC)

Page size on Wikimedia sites is restricted to 2 MB, IIRC. —Simetrical (talk • contribs) 22:03, 25 May 2008 (UTC)

Looks like we might need to hack that down a bit. Stifle (talk) 11:43, 28 May 2008 (UTC)

[edit] Monobook and gadgets not working

I have a number of scripts in my monobook, and have run those successfully for quite a while. My internet died on May 3, and I returned May 24. I noticed fairly quickly the changes to the 'my prefs' menu, and how more of the gadgets are available from there, but all of my scripts have ceased to function completely. I have both purged and bypassed my cache, and have made sure that my browser (Firefox) is Java-compatible, but Twinkle, Friendly, and everything else (including nav popups) no longer work. Does anybody else have this problem, or a solution to it? Prior to my internet going south, everything worked fine. FusionMix 23:27, 27 May 2008 (UTC)

Is JavaScript enabled? Java and JavaScript are two quite different things. —Remember the dot ^(talk) 03:53, 28 May 2008 (UTC)

I checked the 'enable Java' box, but apparently neglected to do the same to the 'enable JavaScript' box. Thanks for the help, that was extremely stupid of me. I feel intelligent now...FusionMix 15:32, 28 May 2008 (UTC)

[edit] Puzzling ref-breaking problem at Manhattan

At the article Manhattan, an extra semicolon is added to a particular instance of "</ref>" making it "<;/ref> and therefore breaking the references from that point on. It has happened in this edit (note the previous "undone" edit which did not touch the politics section), and this edit. Both users who made the edits are in good standing, and they would have no reason to do such a thing. Why is a semicolon being consistently added in that part of the article? Graham87 07:50, 28 May 2008 (UTC)

It's gone now. My guess is that it was just an enduring typo. FusionMix 15:38, 28 May 2008 (UTC)

It's gone because I removed the semicolon, and I remove it in March as well, among other edits. I obviously didn't explain myself clearly ... something on the site is randomly changing a particular instance of </ref> to <;/ref>., therefore breaking the reference system of the article. Why would such a thing happen? I guess it's my hard luck that I've encountered it twice. Graham87 00:50, 29 May 2008 (UTC)

The semicolons appear to have been added in this edit and this edit. -- Boracay Bill (talk) 01:23, 29 May 2008 (UTC)

[edit] Image problem

Hi. Why doesn't this:

Image:Lemniscatic elliptic function.pdf

give me a nice thumbnail image? Robinh (talk) 10:32, 28 May 2008 (UTC)

It is a PDF? --—— Gadget850 (Ed) ^talk - 10:43, 28 May 2008 (UTC)

There are programs that will generate a thumbnail of a PDF. MediaWiki is not among them. --B (talk) 10:49, 28 May 2008 (UTC)

Thanks guys. Yes it is a PDF. So, given that I want a thumbnail, and the image has nice vectorized fonts and is not a photo (I have a few other similar images of other Weierstrass elliptic functions to upload), what would you say the preferred format would be? jpg looks pretty grotty for this type of image. Would png be better? Robinh (talk) 12:25, 28 May 2008 (UTC)

The best option would be to render your work as an SVG. If you don't have the software to do that, you may be able to ask at Wikipedia:Graphic Lab/Images to improve for help. If it is not possible to render as an SVG, then yes, a PNG would be preferred. --B (talk) 13:50, 28 May 2008 (UTC)

Thanks, B. I can create svg no worries. I'll get cracking tonight... Best wishes, Robinh (talk) 14:11, 28 May 2008 (UTC)

[edit] Single user logon

I don't understand how this works. I just went to the Simple English Wikipedia and tried to login with my account here at en., and I was told that I don't have an account there. Isn't SUL supposed to automatically create an account? Is there something I need to do first? Corvus cornix_talk 20:42, 27 May 2008 (UTC)

Go to Special:MergeAccount and then it will work :) -- Cobi^(t|c|b) 20:44, 27 May 2008 (UTC)

Awesome. Thanks, Cobi.  :) Corvus cornix_talk 20:52, 27 May 2008 (UTC)

Thankyou very much, I had no idea about this.  Asenine  08:19, 29 May 2008 (UTC)

[edit] TOCs for major sections?

I'm keen on TOCs (Tables of Contents) in general, for several usability-related reasons. Unfortunately large articles sometimes wind up with a 2-level or 3-level TOC that's more than a screenful in length. Would it be useful and possible to have a multi-part TOC facility, e.g.

Main TOC:

1. Major section 1
2. Major section 2

etc

and then

Major section 1

1.1 Section 1.1

1.2 Section 1.2

etc.

or

Major section 1

1.1 Section 1.1   1.2 Section 1.2   

etc.? Philcha (talk) 12:12, 28 May 2008 (UTC)

You may find that {{toclimit}} works in some cases, but that's only a partial solution to the problem you describe. --ais523 12:22, 28 May 2008 (UTC)

Surely this is un-needed? Perhaps it is just me, but I would find what you are proposing as a problem, and not the current thing. As it goes, surely you can cope with a 1. before your titles in the TOC? On the other hand... I feel I may have misunderstood the question. ¬¬  Asenine  08:16, 29 May 2008 (UTC)

[edit] Typographical Problems with Wikipedia

I looked through Wikipedia's CSS today. It says that the size for body text is 11 points. I think that we should increase it to 12. I noticed that a line of text displayed at a resolution of 1024 by 768 pixels is about 115 characters long without spaces. Next to an infobox, it's about 77 characters long. The ideal length for a line is about 40 characters. Anything over 75 is really hard to read. The most legible sites either use a larger font size or narrow the text block by placing columns to the side of the text. We should also use a serif typeface like Times New Roman for body text. San-serif type styles like Arial are hard to read and are usually reserved for headings.

On another note, I noticed that superscript numerals used for citations here are set in a scientific style. For example, if I wanted to square x, I would type x². But if I wanted to cite a source for x, I should type x.² In Internet Explorer, the former style increases the leading (line spacing) for a line. If we increased the text size, we could also shrink the superscript numerals. I posted these comments on the talk page for the Manual of Style, but was told that I'd need to post them somewhere else to achieve action. So, here I am.--Hello. I'm new here, but I'm sure I can help out. (talk) 23:33, 28 May 2008 (UTC)

I can only point out an error, as I'm not sure about the typographical stuff which you have mentioned. The way you stated that you cite sources is incorrect, it should be done with <ref> tags surrounding an entry.  Asenine  08:13, 29 May 2008 (UTC)

Wikipedia has several skins available under your preferences, which address your concers. You can also customize your personal CSS through your own monobook.css file. — Edokter • Talk • 11:54, 29 May 2008 (UTC)

I think what you say about serif vs sans serif is true for print, but not for reading on the screen (at least not on my screen!), especially for small fonts and/or low resolutions. --Itub (talk) 13:36, 29 May 2008 (UTC)

You're right. But I still think we'd benefit from increasing the font size by a point and by decreasing the size of superscript numerals. I squared a number above, but it I use a <ref> tag, it is rendered like this: x.^[1] It still changes the spacing.--Hello. I'm new here, but I'm sure I can help out. (talk) 18:27, 29 May 2008 (UTC)

Because Internet Explorer is broken.  :) —Simetrical (talk • contribs) 20:32, 29 May 2008 (UTC)

If you find the text line is too wide (and I agree that 115 is outrageously wide), then don't run your browser full screen, and narrow it to a size you find comfortable.  Randall Bart   Talk  04:54, 30 May 2008 (UTC)

[edit] Parenthesis in the URL

I recently edited the "Split (gymnastics)" article. Later, I googled "splits" and found this link near the top of the results list:

http://en.wikipedia.org/wiki/Split_(gymnastics)

When I follow this link I see an old version of the article, yet when I click on "edit this page" I see my revised code, and then when I click "article" I see the revised page, which my browser shows as being at this url:

http://en.wikipedia.org/wiki/Split_%28gymnastics%29

What's happening here, and what can be done to get the google link to show the revised version?

Lambtron (talk) 13:45, 29 May 2008 (UTC)

This sounds like a caching issue, try a hard refresh. xenocidic (talk) 13:48, 29 May 2008 (UTC)

Thanks, xenocidic, you are right. I suspected something like this but got thrown off when a different browser showed the same problem. It appears that neither IE nor FF are smart enough to realize that these two urls point to the same target, so they cache two different pages. Lambtron (talk) 14:42, 29 May 2008 (UTC)

No problem, happy editing. xenocidic (talk) 14:45, 29 May 2008 (UTC)

[edit] enwiki cookies alone not enough to keep a SUL account logged in?

More information re the previous: using a browser, instead of the bot, and using my main (SUL) account, instead of the bot account, and interactively removing all the other wikipedia.org/wikibooks.org/etc. cookies and only leaving the en.wikipedia.org cookies present, effectively logs me out. On the other hand, removing all but the wikipedia.org cookies leaves me logged in.

If this is the case, this would explain my bot problems.

Can anyone else reproduce this? -- The Anome (talk) 23:11, 29 May 2008 (UTC)

Yes. btw, you can get an account de-unified here. Algebraist 23:18, 29 May 2008 (UTC)

[edit] javascript in posts

This may be the wrong place to ask this (if so, I apologize) but why is it possible to add JavaScript to an article (or, presumably, any other page)? I noticed in a recent IP edit to The Sims 2 that valid JavaScript was posted; the code executed successfully (it was nothing more serious than document.write() adding an image and caption to the page). I'm no expert on JavaScript, but does this not present a security threat? Thanks, Aylad ['ɑɪlæd] 01:01, 30 May 2008 (UTC)

Actually, the attempt to add javascript to the page completely failed (it actually seems to have been intended to add some sort of flash ad banner); the javascript code appears as plain text in reference #1. The image comes from the <gallery> tag just above the javascript. Anomie⚔ 01:37, 30 May 2008 (UTC)

Ah, I see that, now that you point it out. (Don't I feel the fool now...) So the special characters in the code get replaced with HTML entities. I understand, thanks! Aylad ['ɑɪlæd] 01:56, 30 May 2008 (UTC)

[edit] Page move throttle: already implemented?

Is this true? Happy‑melon 16:38, 27 May 2008 (UTC)

Been there for years, yes. --brion (talk) 16:46, 27 May 2008 (UTC)

What's it set to? Because the fact that we get so many pagemove vandals would seem to indicate that it's not low enough :D. Perhaps we need Special:UsersAtTheirPageMoveLimit?? Happy‑melon 17:22, 27 May 2008 (UTC)

See config files, and particularly here. New users (I guess that means non-autoconfirmed? not sure offhand, would have to check) get 2 moves every two minutes; others get 8 moves a minute; sysops, bureaucrats, and bots are not limited. —Simetrical (talk • contribs) 14:20, 28 May 2008 (UTC)

I wonder if there's any chance of getting it lowered a bit? Just reducing the burst size to, say, 2 moves in 15 sec would help, since most of the vandal socks get blocked well within one minute of going active. —Ilmari Karonen (talk) 03:10, 30 May 2008 (UTC)

Get consensus and file a bug report asking for the change. —Simetrical (talk • contribs) 17:58, 30 May 2008 (UTC)

[edit] Information displayed below the edit window

Could someone point me to where the stuff below the edit window is controlled in Mediawiki and help me in filing a request (Bugzilla?) to get more information displayed there, or at least reasons why this might not be possible? Thanks. At the moment, the information displayed seems to be limited to a list of templates transcluded on the page and a list of the hidden categories used. What I would like to see added to this is the behaviour produced by each template, which could be listed next to the template in the handy list under the edit window.

ie. Template X contains Category A and Category B (with piped sort keys S and T), and also calls the DEFAULTSORT (value P) and PAGENAME magic words, and also contains Template W (which would be listed elsewhere in the template list anyway, but it would still be good to have each template list where it comes from). The entry for Template W would also say "transcluded from Template X". If the template behaviour was fully explicated below the edit window, it would be a lot easier to work out some of the effects being produced by those templates. Non-hidden categories can be shown by hitting preview, but for completeness it would be good if all categories on a page were listed also (with the ones produced by a template annotated with the name of the template they are coming from). All magic words in use on a page could also be listed as well, plus anything else that can be difficult to find when looking through a large page of wikitext.

Finally, double instances of a template, category, magic word could be listed as well. In particular, I am thinking here about DEFAULTSORT. Ideally, an error message would show at some stage if a page had more than one DEFAULTSORT on it (only the lower one on the page is actually used). That is really a separate request, but related to the above (the general case would be for conflicts to be listed somewhere - for example, failure to close a formatting or reference tag could be flagged up).

Ideally, I would like:

• (a) Comments and advice on the above.
• (b) Some friendly person to file something at Bugzilla (or point to an existing report).
• (c) A friendly developer to take this on (well, if you don't ask!) - it does seem rather simple - only listing a bit more than what is already displayed about what has been done by the software to actually generate the page (in short, extend the current display to explicate all template behaviour listed by each template, and to list categories, templates, and magic words in use on a page, plus anything else that could be usefully listed - maybe a list of all pages linked from a page?).
• (d) Pointers to existing pages and previous discussions and Mediawiki pages (if relevant).

Can anyone help? Carcharoth (talk) 07:59, 28 May 2008 (UTC)

MZMcBride (talk · contribs) know a great deal about this function. Maybe ask him? MBisanz ^talk 08:07, 28 May 2008 (UTC)

At a guess, I would expect that implementing all of what you've suggested would be a prohibitively high increase in the computational power needed to display each edit page. It would be nice to have a toolserver tool that does something like that, however; and some of what you've suggested (eg links from a page) are already efficiently compiled and could be fairly easily output (you can get LinksFromPage from the API IIRC). Happy‑melon 11:53, 28 May 2008 (UTC)

My guess is that it would double things. But surely if the software is interpreting the templates, categories and sortkeys it sees on a page when generating the display (which the reader sees), how much more effort is involved in also outputting the result in a logical order below an edit window? I don't know, but if the software already has to do something along the lines of "display template X - include Category tag A and Category tag B - display Magic word PAGENAME, and also transclude template W (which is inside template X)", then can't it just do the same thing twice? Admittedly, the sort keys are not needed until someone loads the relevant category page: "piped sort keys S and T" and the "DEFAULTSORT (value P)" (incidentally, the magic word PAGENAME can be used as a pipe sort key, as a DEFAULTSORT value, or just a normal display magic word). What could be done there is to have a tab or alternative display that allows people to see the sort key values - imagine a version of the category pages with the sort key values displayed alongside the title (eg. George W. Bush (piped sort key "George" in Category:Bush family, or George W. Bush (DEFAULTSORT sort key "Bush, George W." in Category:Presidents of the United States, or George W. Bush (DEFAULTSORT sort key "Bush, George W." called by Template:Lifetime used in Category:Presidents of the United States). My point here is that some people want to use sortkeys (recently) and categories (been done for a long time) in templates, and it is sometimes difficult to work out exactly where a category or sortkey has come from if there are lots of templates in an article. Helping people see where stuff has come from can only be good. As for LinksFromPage - you say that can be done from the API. Why then is links to a page (Special:WhatLinksHere) done by the software? The only reason seems to be that the latter (whatlinkshere) has to be done server-side, whereas links from a page can be handled by parsing the display. If you are saying that all the stuff I mentioned can be parsed from the display, why then is a list of templates used in a page provided below the edit window? That is useful, so why not other similar stuff as well? Carcharoth (talk) 12:26, 28 May 2008 (UTC)

The thing is that the parser which renders pages for display works systematically: it goes through the wikicode looking for a template call, which it replaces with the wikicode of the template with parameters replaced. Then it goes back to the beginning and starts again, and keeps looping round until there are no more templates to replace. The thing is, it doesn't keep a record of what it's done at each stage, and it would require a considerable allocation of memory for it to do so. Complicated pages like WP:LOCE/R or even the main page might require hundreds of parser cycles to fully render; and don't forget that we get dozens of edits every minute and thousands of page views. Compiling all that information every time the edit window is brought up, whether or not it's actually used, would be prohibitively expensive even if it only doubled the processing time. However, something like special:ExpandTemplates which performed this more rigorous analysis would be tremendously helpful in the instances where such a detailed summary was really necessary.

As for LinksFromPage, what I meant by the API note was merely that this information is already compiled by the MediaWiki software and stored in the database tables. Both LinksFromPage and LinksToPage are available through the API; it's just that only one (LinksToPage) has an equivalent human-readable special page (Special:WhatLinksHere). Adding Special:LinksFromHere would be a trivial exercise for a bored developer. Happy‑melon 13:31, 28 May 2008 (UTC)

Where are these bored developers found and what bribes do they accept? :-) On a more serious note, how is the template and hidden category list below the edit windows generated? Surely by your argument, it must be prohibitively resource intensive to generate that each time? Which is why I am saying that careful extension of what is displayed there might not be impossible. Carcharoth (talk) 13:51, 28 May 2008 (UTC)

The key thing to remember is that the way things are stored in the database has very little similarity to how they're displayed on a page: a fully rendered article contains information drawn from a wide variety of sources. Any information which can be found by a straightforward search of a database table is going to be relatively efficient, whereas anything that requires additional processing is going to have a high server load. For instance, the list of categories each page is in is generated by a search of the Categorylinks table for all links with a cl_from matching the pageid of the page being rendered. A search of the Page props table with those category names reveals which of them should be considered "hidden". The list of templates found on a page is constructed from the Templatelinks table in a very similar way, and the information about whether or not each template is protected is drawn from the Page restrictions table. As the Categorylinks table contains a field for the sortkey of each category link, displaying the sortkey of each page on the category listing (or even on the article itself) would be trivial in terms of server load; however, the table does not note how the sortkey was generated, so that information could not be efficiently compiled without a massive schema change (the field necessary for generating the "number of page in category" statistics was added last year IIRC - and look how long it's taken to update the database tables). So as I said, some of what you suggest could be implemented very efficiently, and perhaps it should. I'd oppose very much more being added by default to the edit page, because it causes said page to take much longer to load; but more of these statistics should be available through one or more special pages. It's just a case of looking through your wishlist above, browsing the various database tables listed at mw:Manual:Database layout, and working out which ideas translate easily into efficient database queries. Happy‑melon 14:16, 28 May 2008 (UTC)

There should be no performance problems in doing this. The lists you'd have to maintain are trivial in size, a couple of kilobytes maybe. As you say, the information is already available to the parser, it just has to be noted at each step. How things are stored in the database is irrelevant, since the page is being parsed on previews at least, and I imagine it would be no great ordeal to parse it when you first click the edit link as well (if that doesn't already happen).

It might be fairly involved to implement it, however. I'm not going to volunteer. Adjusting defaultsort to complain if it occurs multiple times would be simple, but I don't know if it's a good idea: you might legitimately want to set it in a template and then override it in the page's main text, or another template. —Simetrical (talk • contribs) 14:26, 28 May 2008 (UTC)

Ooh. Good. I should look around a bit more, but this does sound promising. About the category totals updating, I'm still waiting for Category:All non-free media to update. Did some process somewhere just give up on that one? Carcharoth (talk) 14:31, 28 May 2008 (UTC)

512 total: that's a laugh!! I stand corrected on the performance issues: of course, if the page is being parsed anyway, then anything that's a trivial addition to the parser is a trivial addition to the whole process. I'd maintain that doing anything to decrease the efficiency of the pageview parser is unlikely to be popular; given the pageview:pageedit ratio (what is it, 100:1 ? something like that, IIRC) I expect the hypothetical developer who implemented all this would probably write a separate, less efficient but more full-featured, parser to be invoked on editpages and previews. I'd start baking cookies, if I were you, Carcharoth :D. Happy‑melon 14:40, 28 May 2008 (UTC)

The maintenance burden of maintaining two separate code paths to do the same thing would far outweigh any minuscule performance benefit from removing a few array pushes. Keep in mind that the parser isn't invoked for the vast majority of views, only for cache misses. —Simetrical (talk • contribs) 20:26, 29 May 2008 (UTC)

Well obviously that's your call - I don't have to work with it! I think that, all other things being equal, I'd still prefer to have a comprehensive Special:AnalysePage than more tacked onto the bottom of the edit screen, even if the only benefit is that my edit page loads faster. Happy‑melon 13:21, 30 May 2008 (UTC)

[edit] Infobox table help

There's an issue with an important template for another project. IE is rendering the tables drastically different from Firefox, etc. I need a lot of help redesigning it so it will look good in both browsers. I have a sample at User:Potapych/Sandbox that's stretched out to its maximum width (The spaces around the images were supposed to be even, but I chose this example because its one of the few instances where this breaks.) I don't want the last line to wrap, and I would like to get it closer to 300px across. The layout also has to stay the same so I can add it back to the template. Potapych (talk) 18:30, 28 May 2008 (UTC)

It looks the same in FireFox 2.0.0.14 and Internet Explorer 7.0.6001.180000 to me. What is the problem?  Asenine  08:17, 29 May 2008 (UTC)

The two colored rows are too thick in IE. IE also seems to ignore the set width and sizes it to however it wants. I added an extra sample, but the difference is really best seen in an article such as this. I don't know why the real infobox in the article is so much wider than the sample I made. Potapych (talk) 14:53, 29 May 2008 (UTC)

Infobox width is a minimum. Normal text will wrap within the defined width, but images force the box width to expand to fit. --—— Gadget850 (Ed) ^talk - 21:04, 30 May 2008 (UTC)

IE's now rendering tables at different widths, regardless of the text. Someone added new images to the article yesterday and a new problem showed up. I don't know why this table is different from all the other ones, but it would be fine if they all looked like it. (Maybe the row headers should be a bit thinner.) The code in the template is identical until the color and text is selected, so it shouldn't be doing that.Potapych (talk) 22:07, 30 May 2008 (UTC)

ETA, it only does that when the parameter Prepressure is used. It can be found at the end of the template code. For some reason, IE doesn't like it.Potapych (talk) 22:15, 30 May 2008 (UTC)

Is there somewhere else I can ask? I don't know the difference between toccolours and infobox classes, and I don't know where else to look.Potapych (talk) 20:54, 30 May 2008 (UTC)

[edit] Wikipedia images issue

Resolved.

For some reason I am unable to display images this morning (but only on Wikipedia, other sites display them fine). I have tried purging my cache and clearing out all of my private data but to no avail, it still continues. This has only happened this morning - and it may be related to me performing a factory reset on my router, but I do not see why that would affect such a specific site. Strangely, even though all of the images in this image of the problem show no images to be working, the background image and the user image display fine. Any ideas?  Asenine  07:53, 29 May 2008 (UTC)

By the way, I have just tried, and I can see them when I open them directly (ie. via a url, not a rendered image from the MW software). Strange. I think the two images that do display may do it because they are not resized. All images display normally in IE, but I am certainly not using that hunk of junk.  Asenine  07:55, 29 May 2008 (UTC)

I'm assuming you're using Internet explorer? If so, check to see if you have set a preference to block images from upload.wikimedia.org (if IE has this feature). On my firefox, I was able to get the exact scenario in your screen shot by blocking upload.wikimedia.org -- penubag  (talk) 09:17, 29 May 2008 (UTC)

I'm using FireFox 2.0.0.14. As for blocking, it would have to be all sites because I have now noticed it is happening for any image which is resized via code, HTML or otherwise. Not in IE though, they work in IE.  Asenine  09:45, 29 May 2008 (UTC)

I am going to perform a complete hard reinstall of FireFox. I'll update you when it is done.  Asenine  09:46, 29 May 2008 (UTC)

A hard reinstall with 2.0.0.14 or 3 RC1 still gives the same problems. That said, I think I may have failed to hard install - I might have missed some registry entries as it still had my bookmarks. Everything looks fine in Opera (which I am using now), but this browser is much too complicated to customize for my liking.  Asenine  14:56, 29 May 2008 (UTC)

It is possible that you have some ad blocker that is blocking the images. Check you computer for any software that may do this. But that being said, you claim that other browsers aren't blocking the images so most likely the problem is Firefox. Have you tried Add/Remove utility under the control panel to remove firefox (if you're using a PC)? -- penubag  (talk) 06:41, 30 May 2008 (UTC)

As I said before, I did a hard reinstall (ie. uninstall and then remove what the uninstaller leaves behind). I did have an adblocker on, but I disabled it and it still happens. The newly installed one does not have an ad blocker installed, but it is still doing it. 15:59, 30 May 2008 (UTC)

Some broken anti-virus/firewall software tries to block ads, I believe. You could check for that. —Simetrical (talk • contribs) 18:00, 30 May 2008 (UTC)

Turns out it was something to do with a plugin for FireFox. Not sure which one, but I discovered this evening that my hard reinstall actually missed the main folder, and when I deleted that and installed everything was back to normal. Thanks for the help, everyone. :)  Asenine  22:24, 30 May 2008 (UTC)

[edit] Editing Login Cookies

When I tried to edit the expire date of the login cookies (usually 30 days from login) I experienced problems with some common Firefox-add-ons. For example, the tool Edit Cookies displays "NaN" instead of the current date in the "New Expire Date" fields (at least under Windows 2000; with Iceweasel/Debian it displays the current date correctly). But much worse, any edit will be forgotten after leaving the edit dialog. Cookies from other website (including Bugzilla) can be edited without problems. Could this be a bug or an intentional prohibition of editing these cookies? However, via Opera you can edit the Wikipedia cookies without any problems (and without the need of add-ons), therefore I would guess it's a but, not a feature.--SiriusB (talk) 13:47, 29 May 2008 (UTC)

You might check whether your cookie-editing tools have a known problem working with cookies marked HttpOnly, as our session & token cookies currently are. --brion (talk) 15:31, 29 May 2008 (UTC)

Are you sure the syntax is right? NaN means 'not a number'.  Asenine  23:23, 29 May 2008 (UTC)

Either the syntax of the cookies is buggy or the cookie editor itself, because the NaNs appeared even before my editing trials. But only in Windows; under Linux the date is shown correctly but cannot be edited.--SiriusB (talk) 07:31, 30 May 2008 (UTC)

[edit] Single sign-on breaks old bot code

Unifying my bot account has completely broken my bot's login code. I can see how to fix it, but the fix is rather involved, since I would have to completely emulate the cascade of image-load and cookie operations associated with a new-style SSO login. Would it be possible for the SSO mechanism to still be able to recognize just the login cookies in the original domain I am logging into, even when an account is unified for single sign-on and the cross-domain cookies are missing? Otherwise, I'm going to either have to write a load of SSO bot-login code, or have to port it to a framework which already has this implemented. Either of these would be painful, and I'd prefer not to have to do either. -- The Anome (talk) 20:41, 29 May 2008 (UTC)

Alternatively, is it possible to dis-unify an account? -- The Anome (talk) 20:46, 29 May 2008 (UTC)

You shouldn't need to do any image loading. Just accept the cookies you get on that site you log in on and pass them back for subsequent requests. --brion (talk) 20:48, 29 May 2008 (UTC)

I'm setting enwiki_session, enwikiUserID and enwikiUserName, which are the three cookies I get back from the login on enwiki. Are there any others I should be setting? -- The Anome (talk) 21:12, 29 May 2008 (UTC)

You should have centralauth_User and centralauth_Session cookies in there as well (and optionally centralauth_Token). Note they'll be set for the parent domain, eg "wikipedia.org" here. --brion (talk) 00:02, 30 May 2008 (UTC)

Thanks. I'll take a closer look at my cookie-handling code to see how it handles setting cookies in parent domains. -- The Anome (talk) 00:25, 30 May 2008 (UTC)

Update: yes, you're quite right, the cookies are all sent in the single HTML transaction. It's clearly a bug in my cookie store code. Thanks. -- The Anome (talk) 00:54, 31 May 2008 (UTC)

[edit] Renaming ro.wikipedia.org into mo-ro.wikipedia.org (Moldovan-Romanian languages)

I would like to ask first of all whether this is the appropriate place on Wikipedia to start this discussion and if it is not, could someone please indicate where should one start it? Thanks--Moldopodo (talk) 01:07, 30 May 2008 (UTC)

This should be discussed on Meta-Wiki. [8] -- penubag  (talk) 06:31, 30 May 2008 (UTC)

You may want to start by asking at the language subcommittee's talk page (and while you are at it, read the Moldovan language project section). — Note that a Moldovan Wikipedia still exists, although it is locked (it can be read but not edited) following the proposal for closing it. - Regards, Ev (talk) 07:08, 30 May 2008 (UTC)

Thank you so much, I have just created the request on Metawiki[9]--Moldopodo (talk) 11:43, 30 May 2008 (UTC)

[edit] Login problems

Did something change in the past few hours with respect to logging in and accessing the API? My enhanced watchlist program stopped working, and is now getting an error about not being logged in when it tries to query api.php. --Carnildo (talk) 08:57, 30 May 2008 (UTC)

You haven't just activated SUL, have you? If so, the threads above might be relevant. Algebraist 10:38, 30 May 2008 (UTC)

I activated it several months ago, and haven't had a problem until now. --Carnildo (talk) 20:26, 30 May 2008 (UTC)

[edit] Customizing a toolbox

Hi to anyone who could help!
How can I customize the “toolbox” we have just below the edit window. I do not need Cyrillic, Greek and IPA scripts and most of the Latin letters (I have non-virtual keyboards for this), but need e.g. mkhedruli alphabet and, most of all, those </br>, <small></small> and others.
Furthermore, is it possible to add to this toolbox templates and some infoboxes I often use?
Many thanks forward for helping! ✓ Kanġi Oĥanko (talk) 15:53, 30 May 2008 (UTC)

Go to the template you are embedding and copy the code into somewhere in your userspace. From there you can remove parts of the code and add too.  Asenine  15:56, 30 May 2008 (UTC)

What about the "other alphabets" ? xenocidic (talk) 16:01, 30 May 2008 (UTC)

[edit] Did I find a bug in the history?

Someone put something on this page, but the history shows no contribution [10]. They did not sign it either. I am mystified. Is this a bug? I have never seen this before.--Filll (talk | wpc) 17:52, 30 May 2008 (UTC)

Sometimes, you are just seeing a cached version of a page and have to wait a minute or hit the "refresh" button on your browser to get it to update. I see two contributions when I look at the history you linked. --B (talk) 18:01, 30 May 2008 (UTC)

Wow that is sort of surprising to me. But yes when I refreshed it did show that two had contributed. Even though my post to that page was a long time ago, so there should have been plenty of time for the cache etc to be updated. I feel like a moron.--Filll (talk | wpc) 18:18, 30 May 2008 (UTC)

Something is definitely wrong here. Refresh [11]. Sometimes the oldest two edits appear, sometimes they do not. Same is true for [12]. --- RockMFR 18:44, 30 May 2008 (UTC)

This is happening everywhere. Seems to be affecting revisions made around 2008-05-25 and earlier on any recently created pages. --- RockMFR 19:16, 30 May 2008 (UTC)

I wouldn't call it wrong. I often see messages saying something like "this list may not include revisions in the last xx seconds" (minutes or hours?). I don't know how the servers work but I think you are just getting the history some of the time from a server that hasn't been updated and some of the time from one that has. 199.125.109.104 (talk) 19:21, 30 May 2008 (UTC)

Some more examples: [13] [14] [15] --- RockMFR 19:23, 30 May 2008 (UTC)

Domas is doing load testing with a MySQL 5.0 server, but it seems to have a damaged data set, which is causing these issues. We're working on it. --brion (talk) 19:34, 30 May 2008 (UTC)

Ok, it's de-pooled and things look stable again. --brion (talk) 19:38, 30 May 2008 (UTC)

[edit] Overlapping letters with HTML or LaTeX

Image:NLC-STL-Insignia.png I am told in Would a text-based recreation of a logo be illegal? in the Policy portion of the Pump that a text version of the image shown at the right would be legal in a user box. So how do I do that? I don't care if it looks exact. If it has red letters on a white background or visa versa with the letters in the correct positions, I will be happy. Will (Talk - contribs) 00:11, 29 May 2008 (UTC)

Only way I can think of doing that would be by using the position:absolute div style. But that would be downright awkward and impractical. It also, now I think about it, probably wouldn't work. I am still posting this, as it might give someone else an idea of what to do with the div tag. Also, remember that what you read there has not been through any sort of consensus.  Asenine  08:03, 29 May 2008 (UTC)

If you could upload the text as images, Template:Superimpose will work. -- penubag  (talk) 08:20, 29 May 2008 (UTC)

Anything that would be legal if you uploaded it as text encoded in HTML would be equally legal if you uploaded it as text encoded in SVG. SVG easily allows overlapping letters. For that matter, anything legal in either one would be legal if rasterized to PNG.

On the other hand, the text would also not accurately represent the logo of the team, and so would probably not be useful for encyclopedic purposes. Misrepresenting the logo of a team may not be a copyright violation, but I wouldn't be surprised if it violated trademark law (since it would be a misrepresentation), and anyway it doesn't serve an encyclopedic purpose since it misinforms readers. Moreover, while text alone is not copyrightable, precisely-positioned and -colored text could reach the threshold of copyrightability. In any event you almost certainly shouldn't try to dodge the copyright policy this way; accept the fact that the logo will only be present on a few key pages, don't try to make up pseudo-logos. But that's a separate issue. —Simetrical (talk • contribs) 20:41, 29 May 2008 (UTC)

I can't use the actual image as I want to use it in a user box. Also, {{Superimpose}} won't help as the text is text. Will (Talk - contribs) 01:48, 31 May 2008 (UTC)

Yes, you can upload individual type-faces as images licensed as {{PD-font}} and superimpose them over each other to make a composite of what you want. -- penubag  (talk) 05:03, 31 May 2008 (UTC)

Well, I created this image, but I question that {{PD-font}} could be used with it as the image wasn't created to demo characters. Rather, it is meant to reproduce a logo. Please note that the official logo has characters whose outlines are merged — not overlapping. Will (Talk - contribs) 05:53, 1 June 2008 (UTC)

[edit] I added a track map and coodinates to Pacific Raceways but they won't appear

Resolved. My mistake, although I don't know why the image is now showing up. Will (Talk - contribs) 06:04, 31 May 2008 (UTC)

Why not? What's wrong? Will (Talk - contribs) 03:06, 31 May 2008 (UTC)

Well, Drdisque appears to have fixed it. I seem to have put a pipe before the call to {{Coord}}. I still don't know why the image didn't appear. Drdisque didn't change that. But now the image appears. Will (Talk - contribs) 06:04, 31 May 2008 (UTC)

The image didn't appear because there was no pipe immediately before 'image='. I've changed the documentation to make this error impossible. Algebraist 17:00, 31 May 2008 (UTC)