Wikipedia:Village pump (technical)/Archive 69

• Village pump
  • Policy
  • Technical
  • Proposals (persistent)
  • Idea lab
  • WMF
  • Miscellaneous

Village pump (technical) archive  

This page contains discussions that have been archived from Village pump (technical).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

---

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

Creating anchors

I tried to use {{anchor}} and found the documentation unhelpful. Any recommendations for how to use {{anchor}}, or whether using html <span id="MYANCHOR"/> is better? Please comment at Template talk:Anchor#Usage is confusing (wrong?). Johnuniq (talk) 08:54, 23 December 2009 (UTC)

<span/> is invalid and is converted to <span></span> when it gets run through HTML Tidy. Putting {{anchor}} or <span></span> in the header have the issue of usually screwing up edit summaries. --MZMcBride (talk) 09:07, 23 December 2009 (UTC)

MZMcBride: No, using <span> tags doesn't screw up the edit summaries. So using <span> tags directly is the solution. But please, let's keep this discussion over at Template talk:Anchor#Usage is confusing (wrong?).

--David Göthberg (talk) 15:10, 23 December 2009 (UTC)

{{dts}} not appearing

Every National Register of Historic Places list has a dates column, using {{dts}}, for when places were put on the Register. For the past few hours, this template has been displaying exactly nothing: as you can see at National Register of Historic Places listings in Beaver County, Pennsylvania, the "Date listed" column appears to be blank. The template hasn't been vandalised, the code still recognizes it (try sorting the column for proof), and it likewise fails to appear when I try to put the template in non-table prose text in other articles. Any idea what's wrong? Nyttend (talk) 23:08, 23 December 2009 (UTC)

I see the first row with the date of October 24, 1996. ---— Gadget850 (Ed) ^talk 23:12, 23 December 2009 (UTC)

{{Dts/fmt}} was vandalized by an anonymous editor. It should be fixed now. Svick (talk) 23:18, 23 December 2009 (UTC)

That's a pretty insightful vandal, figuring out how to vandalise a semiprotected page while not logged in and making it seem that the page wasn't vandalised. Thanks for the help; I've never quite figured out much in the template namespace. I've semiprotected that template and the one to which it's a redirect. Nyttend (talk)

In that case, you might want to semiprotect other subpages of that template (except /doc and /sandbox*). Svick (talk) 23:32, 23 December 2009 (UTC)

All done; thanks for the pointer. Nyttend (talk) 23:49, 23 December 2009 (UTC)

Secure server

Currently most links to other Wikimedia projects (like Wiktionary) point to the normal servers, even when the user is using the secure server. I am now updating the links in the MediaWiki interface and in other places such as the Main Page and the sister project templates. I make it so users on the secure server see secure links, while users on the normal servers see normal links.

This means I am editing many high-visibility places, and that I am doing a site wide change, so I am announcing this in case anyone has any comments about it. See Wikipedia talk:Secure server#Sister project links for more on this and to discuss it.

--David Göthberg (talk) 01:25, 24 December 2009 (UTC)

Embedding Google Street View images onto Wikipedia pages

Google Street View is a great source for images of many places. It is a way to obtain images without having to physically go to these places camera-ready, something that is not always easy to do.

Google indeed authorizes their maps and street view images to be used on one's websites. But this is under the condition that the HTML be embedded straight onto the site and that a screenshot not be uploaded. You can read about this in detail on this page where it says:

If you want to use Content from Google Maps, Google Earth, Street View, etc. on your website, embed it within the site rather than uploading screenshots. This means the Content will be loaded directly from Google's servers, and will automatically have appropriate attribution.

Now, is there a way that Google's images can be added to Wikipedia pages in a manner that complies with this? If not, can Wikipedia be made so this is possible? Sebwite (talk) 01:55, 24 December 2009 (UTC)

Because the images must remain on Google's servers, this effectively is saying that derivative works are prohibited; all images here, other than fair-use, must permit derivative works, so I doubt that the technical side of doing this really matters. Nyttend (talk) 05:43, 24 December 2009 (UTC)

asae38hfhue

So, I noticed some characters flickering up as page source is rendered some time back, notably at the end of each blank line of long templates. Eventually I managed to capture them, they can be Googled to show that they have a strong link with WP but that's about it. Any ideas? Rich Farmbrough, 21:43, 20 December 2009 (UTC).

Can you be more specific? Where exactly are they appearing? Isn't the whole string similar to UNIQ28cf615b3f10792f-nowiki-0000000B-QINU? If yes, that would mean they are strip markers (whatever that means), see T19329. Svick (talk) 22:22, 20 December 2009 (UTC)

@Rich: Huh? First you insert these characters [1], a day later you remove them again [2], and then nearly three weeks later you ask about their meaning here? (This article is the only pertinent Google hit for this character sequence, and then only on some external site that somehow has the first diff as an HTML page.) What's up? Lupo 09:09, 24 December 2009 (UTC)

Wikipedia Mobile

The mobile survey that Wikipedia posted on Google Docs is having technical issues and I was unable to successfully submit a survey. For this reason I have copied and pasted the 'comments" section from the survey onto here.

I would like to point out that the mobile site has a problem on my second generation IPod touch (using Safari). When i open an article and rotate my device horizontal from the vertical position, a third of the screen stays the way it was before, also if i click a link to a new article while in this horizontal position, that part of the screen, again, freezes and stays on the previous article as if the screen was still vertical. i find myself clicking "view this page on regular Wikipedia" for almost every article (or fixing the problem by rotating my IPod touch vertical and then back horizontal again.) one more thing, i was UNABLE to do this survey on my device (didn't respond after "submit" ) and am currently doing it on my desktop. I love you guys and will donate soon when i set up paypal! keep the ads away!

Thanks, Danny —Preceding unsigned comment added by Dninyo (talk • contribs) 20:58, 23 December 2009 (UTC)

I have relayed this to the developer of the mobile interface. —TheDJ (talk • contribs) 10:03, 24 December 2009 (UTC)

Some trouble with multi-prefix search

I've installed a multi-prefix search form at WT:ROMANIA to search through the previously fragmented notice boards. But it doesn't seem to work quite right. For instance searching for abroad returns a bunch of pages, but searching for Badea (a Romania name, but not using any diacritics) returns nothing, even though it definitely appears on some of those pages, e.g. here. Does anyone know what might cause this? Pcap ping 02:17, 24 December 2009 (UTC)

Your example Wikipedia:Romanian Wikipedians' notice board/Archive 11 was created 8 hours ago and has not been indexed by the search function yet so neither abroad nor Badea is found there by searching. Do you know of other pages where Badea occurs? PrimeHunter (talk) 02:40, 24 December 2009 (UTC)

Don't think so. You're right that Badea is found at the previous location [3] of that contents before I archived it. Pcap ping 02:42, 24 December 2009 (UTC)

However, that's not the only anomaly. The word "checkuser" is found in archives 8 and 10, but not in 9, where it also appears. Kinda strange. I'll try again in some 24 hours see if the indexing gets sorted out. Pcap ping 02:47, 24 December 2009 (UTC)

This is the same issue. "checkuser" is found in two old pages but not in a page created 11 hours ago. Note that Archive10 in the search is the old Wikipedia:Romanian Wikipedians' notice board/Archive10 while archive 9 in your link is the new talk archive Wikipedia talk:Romanian Wikipedians' notice board/Archive 9. PrimeHunter (talk) 03:07, 24 December 2009 (UTC)

Pohta, PrimeHunter is right that it seems the pages in question have not yet been indexed, but also, you need to search the main noticeboards as well as the archives. You can do this by specifying the prefix without the ending "/". Example: a prefix of "User talk:Stmrlbs" will search all pages starting with "User talk:Stmrlbs" - this means the main talk pages and all talk page archives because the archivs start with "User talk:Stmrlbs/" - which means they also start with "User talk:Stmrlbs". However if you specify a search prefix of "User talk:Stmrlbs/", then you will only search the archives, because the main talk page does not begin with "User talk:Stmrlbs/". I have changed the search prefixes on WT:Romania to inlude the main pages. Please check to make sure that this is what you want. The problem is now you will find "Badea"... on the main noticeboard page, because that is where it was when it was last indexed. And.. unfortunately, there is no software function to tell you if something was on a main page on a certain date, then what archive is it in now? For that matter, there is no housekeeping function in Wikipedia to update links going to text that has been archived.. which, imo, is a real weakness in wikipedia. But.. that's the way it is. Sorry. stmrlbs|talk 04:11, 24 December 2009 (UTC)

It looks like the indexing kicked in, and the search works correctly now. Thank you both for your help. Pcap ping 19:48, 24 December 2009 (UTC)

Odd Firefox text search problem

Moved from Wikipedia talk:Village pump Svick (talk) 00:53, 25 December 2009 (UTC)

I seem to be having an odd problem recently searching for text within English-language Wikipedia pages on a Firefox browser. For example, within Andalusian independentist conspiracy (1641), it will not find Portugal. Doesn't seem to be a problem outside Wikipedia of even in es-wiki. Closed & re-opened browser, still having the problem. It might imaginably be something about a widget I use, but I haven't knowingly added any lately. Has anyone else experienced something similar? Any insights would be welcome. - Jmabel | Talk 00:46, 25 December 2009 (UTC)

Are you sure you don't have case sensitivity on and searching for “portugal” (note the small p)? It works okay for me in Firefox on that article. Svick (talk) 00:56, 25 December 2009 (UTC)

That turned out to be precisely it. I was just coming here to say I'd worked it out, but your comment preceded me. - Jmabel | Talk 01:30, 25 December 2009 (UTC)

Image not visible

In trigonometric functions, I see the caption that accompanies this image, but I don't see the image itself. I do see the other image right above it. What's going on? Michael Hardy (talk) 17:35, 22 December 2009 (UTC)

....and now I find that when I add |250px here, then I can't see the image on this page, whereas I can when it's all alone, as above. The image above it in trigonometric functions also has that size restriction, but no such problem afflicts it. Michael Hardy (talk) 17:37, 22 December 2009 (UTC)

Perhaps your browser does not support .svg or .png? What browser/OS are you using? Intelligentsium 17:41, 22 December 2009 (UTC)

See instructions on purging images to solve this problem. I followed the steps at that page for your image, but I can't tell whether they worked or not because, as it says on my user page, I'm blind. Does the image work for you now? Graham87 06:05, 23 December 2009 (UTC)

It's still not working.

"Intelligentsium", does it work for you? If my browser fails to support .svg or .png, then why is the other illustration on that page, right above the one I pointed out, working normally for me? Michael Hardy (talk) 00:56, 24 December 2009 (UTC)

It's working for me, have you tried bypassing your cache? Svick (talk) 01:00, 24 December 2009 (UTC)

Just tried that. On the Google Chrome browser. Still not working. And the other image, just above it, still works normally. I've tried this one three different browsers now: two on my home computer, one on an on-campus linux machine with seamonkey. It fails on all of them. Michael Hardy (talk) 01:52, 24 December 2009 (UTC)

For whatever it's worth, I can see it using Firefox 3. Pcap ping 02:22, 24 December 2009 (UTC)

Do you have any ad blocking software that could be blocking that image in particular because of some text in its URL? Maybe this isn't very likely because you've attempted to view the image on two different machines, but it's worth a try. Graham87 02:43, 25 December 2009 (UTC)

Not that I'm aware of. Michael Hardy (talk) 21:41, 25 December 2009 (UTC)

Formatting problem with bulleted lists by floaters?

Is there some more elegant way to deal with the following problem I have discovered with bulleted and numbered lists next to floating images and tables? (If you are using a very wide-screen monitor, I suggest varying the width of the browser window so you can see how the text moves around in relation to the floating table.)

---

Example floater table, align="left"
Example floater data
Example floater data
Example floater data
Example floater data
Example floater data
Example floater data
Example floater data

Normal unbulleted text has correct spacing next to table. Normal unbulleted text has correct spacing next to table. Normal unbulleted text has correct spacing next to table. Normal unbulleted text has correct spacing next to table. Normal unbulleted text has correct spacing next to table. Normal unbulleted text has correct spacing next to table. Normal unbulleted text has correct spacing next to table.

• Bulleted Item for some reason is "outdented" next to floating table, in both Firefox and Internet Explorer, which looks downright wrong but there is no obvious solution as to whatever is causing this bizarre bullet-item formatting problem that is screwing up the page layout and document flow. (If you are using a very wide-screen monitor I suggest varying the width of the browser window so you can see how this text moves around in relation to the floater.)

• Bulleted Item for some reason is "outdented" next to floating table, in both Firefox and Internet Explorer, which looks downright wrong but there is no obvious solution as to whatever is causing this bizarre bullet-item formatting problem that is screwing up the page layout and document flow. (If you are using a very wide-screen monitor I suggest varying the width of the browser window so you can see how this text moves around in relation to the floater.)

• Bulleted Item for some reason is "outdented" next to floating table, in both Firefox and Internet Explorer, which looks downright wrong but there is no obvious solution as to whatever is causing this bizarre bullet-item formatting problem that is screwing up the page layout and document flow. (If you are using a very wide-screen monitor I suggest varying the width of the browser window so you can see how this text moves around in relation to the floater.)

Normal unbulleted text has correct spacing next to table. Normal unbulleted text has correct spacing next to table. Normal unbulleted text has correct spacing next to table. Normal unbulleted text has correct spacing next to table. Normal unbulleted text has correct spacing next to table. Normal unbulleted text has correct spacing next to table.

---

A partial-fix (hack) I have discovered is to enclose the bulleted list in an extremely simple one-cell non-float table, such as:
{|
|-
|
* Item 1
* Item 2
* Item 3
|}
En-tabling the bulleted list fixes the page layout...... but by putting the entire list in a single large table, large bulleted paragraphs don't cleanly move to the left side of the page below a floating table:

---

Example floater table, align="left"
Example floater data
Example floater data
Example floater data
Example floater data
Large, ugly, open whitespace
below this floater due to a
single large table enclosing
the bulleted list. Sigh.

Normal unbulleted text has correct spacing next to table. Normal unbulleted text has correct spacing next to table. Normal unbulleted text has correct spacing next to table. Normal unbulleted text has correct spacing next to table. Normal unbulleted text has correct spacing next to table. Normal unbulleted text has correct spacing next to table. Normal unbulleted text has correct spacing next to table.

Bulleted Item for some reason is "outdented" next to floating table, in both Firefox and Internet Explorer, which looks downright wrong but there is no obvious solution as to whatever is causing this bizarre bullet-item formatting problem that is screwing up the page layout and document flow. (If you are using a very wide-screen monitor I suggest varying the width of the browser window so you can see how this text moves around in relation to the floater.) Bulleted Item for some reason is "outdented" next to floating table, in both Firefox and Internet Explorer, which looks downright wrong but there is no obvious solution as to whatever is causing this bizarre bullet-item formatting problem that is screwing up the page layout and document flow. (If you are using a very wide-screen monitor I suggest varying the width of the browser window so you can see how this text moves around in relation to the floater.) Bulleted Item for some reason is "outdented" next to floating table, in both Firefox and Internet Explorer, which looks downright wrong but there is no obvious solution as to whatever is causing this bizarre bullet-item formatting problem that is screwing up the page layout and document flow. (If you are using a very wide-screen monitor I suggest varying the width of the browser window so you can see how this text moves around in relation to the floater.) Bulleted Item for some reason is "outdented" next to floating table, in both Firefox and Internet Explorer, which looks downright wrong but there is no obvious solution as to whatever is causing this bizarre bullet-item formatting problem that is screwing up the page layout and document flow. (If you are using a very wide-screen monitor I suggest varying the width of the browser window so you can see how this text moves around in relation to the floater.)

Normal unbulleted text has correct spacing next to table. Normal unbulleted text has correct spacing next to table. Normal unbulleted text has correct spacing next to table. Normal unbulleted text has correct spacing next to table. Normal unbulleted text has correct spacing next to table. Normal unbulleted text has correct spacing next to table.

---

The more complete fix (hack on a hack) I have discovered is to enclose each item in the bulleted list in its own one-cell non-float table, such as this:
{|
|-
|
* Item 1
|}
{|
|-
|
* Item 2
|}
{|
|-
|
* Item 3
|}

Bulleted items below the floating table/image now move out to the side of the browser window correctly, and whitespace is minimized:

---

Example floater table, align="left"
Example floater data
Example floater data
Example floater data
Whitespace below floater
is minimized, by putting
each bullet in its own
1-cell non-float table.

Normal unbulleted text has correct spacing next to table. Normal unbulleted text has correct spacing next to table. Normal unbulleted text has correct spacing next to table. Normal unbulleted text has correct spacing next to table. Normal unbulleted text has correct spacing next to table. Normal unbulleted text has correct spacing next to table. Normal unbulleted text has correct spacing next to table.

Bulleted Item for some reason is "outdented" next to floating table, in both Firefox and Internet Explorer, which looks downright wrong but there is no obvious solution as to whatever is causing this bizarre bullet-item formatting problem that is screwing up the page layout and document flow. (If you are using a very wide-screen monitor I suggest varying the width of the browser window so you can see how this text moves around in relation to the floater.)

Bulleted Item for some reason is "outdented" next to floating table, in both Firefox and Internet Explorer, which looks downright wrong but there is no obvious solution as to whatever is causing this bizarre bullet-item formatting problem that is screwing up the page layout and document flow. (If you are using a very wide-screen monitor I suggest varying the width of the browser window so you can see how this text moves around in relation to the floater.)

Bulleted Item for some reason is "outdented" next to floating table, in both Firefox and Internet Explorer, which looks downright wrong but there is no obvious solution as to whatever is causing this bizarre bullet-item formatting problem that is screwing up the page layout and document flow. (If you are using a very wide-screen monitor I suggest varying the width of the browser window so you can see how this text moves around in relation to the floater.)

Bulleted Item for some reason is "outdented" next to floating table, in both Firefox and Internet Explorer, which looks downright wrong but there is no obvious solution as to whatever is causing this bizarre bullet-item formatting problem that is screwing up the page layout and document flow. (If you are using a very wide-screen monitor I suggest varying the width of the browser window so you can see how this text moves around in relation to the floater.)

Normal unbulleted text has correct spacing next to table. Normal unbulleted text has correct spacing next to table. Normal unbulleted text has correct spacing next to table. Normal unbulleted text has correct spacing next to table. Normal unbulleted text has correct spacing next to table. Normal unbulleted text has correct spacing next to table.

---

So as you can see I have found a solution, but this hackery seems unnecessary on the part of article editors to make bulleted lists and floating tables/images work together cleanly. Also I don't feel comfortable leaving weird stuff like this in an article without some additional <!-- hidden text to explain to other editors what the heck is going on here -->.

Is there some better way to deal with the bulleted-list formatting by floaters? Is there something that can be done to the back-end code to fix this, so I don't need to manually hack page formatting like this? DMahalko (talk) 15:44, 23 December 2009 (UTC)

This is standard CSS behaviour with floating blocks and inline text (and list items bullets are normal inline elements). And yes, it's horrendous behavior, and the guys in the CSS standards group better be very sorry about it, but that does not change anything. Your analysis was accurate and there is no other way to solve this problem as far as I know. —TheDJ (talk • contribs) 15:53, 23 December 2009 (UTC)

Since the page layout markup for lists are not directly created by editors, but instead use wikimedia's simplified markup, it appears possible for this CSS page rendering error to be "fixed" in the page rendering code...

Rather than:

* item foo

* item bar

Becoming:

<ul><li>item foo</li></ul>

<ul><li>item bar</li></ul>

It could be marked up as:

<table><tr><td><ul><li>item foo</li></ul></td></tr></table><br>

<table><tr><td><ul><li>item bar</li></ul></td></tr></table><br>

DMahalko (talk) 03:22, 24 December 2009 (UTC)

I give very little chances to a developer approving this. (Hacks like this are heavily frowned upon) —TheDJ (talk • contribs) 09:53, 24 December 2009 (UTC)

By appling a fix like this at the wikimedia code level, it at least allows for browser detection code so that if the CSS problem is fixed eventually, then the wikimedia software could check the browser version and not apply this hack to browsers newer than the current ones that (hopefully) won't need the hack.

If this can't/won't be approved at the developer level, I have no problem with manually applying this table-hack to every wrongly-formatted bulleted-list I can find....... hmmm, I see an incorrectly formatted bulleted list on TheDJ's user-page.. ;)

So just to be clear, nobody else cares if I am silently applying this table-fix/hack all over the place to every incorrectly-aligned bulleted list in every article, without any way to track the table-hack applications, and with no way to be forward-looking to an eventual future browser CSS fix?

(I suppose I could make a hidden category to track this so they can be removed someday if browsers ever do handle lists and floating tables correctly.)

DMahalko (talk) 23:36, 24 December 2009 (UTC)

Simply put, don't do this in articles. Perhaps an easier way to deal with this would be to float things right instead of left in problem places?... --Izno (talk) 23:52, 24 December 2009 (UTC)

Your suggestion seems unacceptable because it greatly limits the ability for editors to format pages that are dense with floating objects like the TOC, infoboxes, images, and various floating objects.

Many articles are way overloaded with stuff all along the right side that look horribly ugly and run float-images way out of alignment with the text, unless Template:Clear is used all over the place.

Requiring editors to force things to be right-justified just to make bulleted lists not look ugly is a severe editing and layout limitation, if there is this simple fix available for the problem. DMahalko (talk) 00:09, 25 December 2009 (UTC)

Templates to fix bullet alignment / tracking category

I suppose I could make templates which include a hidden tracking category within. I am not a programmer so these templates are probably not created correctly:

• Template:Indent-fix-begin
• Template:Indent-fix-end
• Hidden maintenance category: Category:List alignment fix next to floating tables

And how well does it work? Testing:

---

Example floater table, align="left"
Example floater data
Example floater data
Example floater data
Whitespace below floater
is minimized, by putting
each bullet in its own
1-cell non-float table.

Normal unbulleted text has correct spacing next to table. Normal unbulleted text has correct spacing next to table. Normal unbulleted text has correct spacing next to table. Normal unbulleted text has correct spacing next to table. Normal unbulleted text has correct spacing next to table. Normal unbulleted text has correct spacing next to table. Normal unbulleted text has correct spacing next to table.

{{Indent-fix-begin}}

• Bulleted Item for some reason is "outdented" next to floating table, in both Firefox and Internet Explorer, which looks downright wrong but there is no obvious solution as to whatever is causing this bizarre bullet-item formatting problem that is screwing up the page layout and document flow. (If you are using a very wide-screen monitor I suggest varying the width of the browser window so you can see how this text moves around in relation to the floater.)

{{Indent-fix-end}} {{Indent-fix-begin}}

• Bulleted Item for some reason is "outdented" next to floating table, in both Firefox and Internet Explorer, which looks downright wrong but there is no obvious solution as to whatever is causing this bizarre bullet-item formatting problem that is screwing up the page layout and document flow. (If you are using a very wide-screen monitor I suggest varying the width of the browser window so you can see how this text moves around in relation to the floater.)

{{Indent-fix-end}} {{Indent-fix-begin}}

• Bulleted Item for some reason is "outdented" next to floating table, in both Firefox and Internet Explorer, which looks downright wrong but there is no obvious solution as to whatever is causing this bizarre bullet-item formatting problem that is screwing up the page layout and document flow. (If you are using a very wide-screen monitor I suggest varying the width of the browser window so you can see how this text moves around in relation to the floater.)

{{Indent-fix-end}} {{Indent-fix-begin}}

• Bulleted Item for some reason is "outdented" next to floating table, in both Firefox and Internet Explorer, which looks downright wrong but there is no obvious solution as to whatever is causing this bizarre bullet-item formatting problem that is screwing up the page layout and document flow. (If you are using a very wide-screen monitor I suggest varying the width of the browser window so you can see how this text moves around in relation to the floater.)

{{Indent-fix-end}}

Normal unbulleted text has correct spacing next to table. Normal unbulleted text has correct spacing next to table. Normal unbulleted text has correct spacing next to table. Normal unbulleted text has correct spacing next to table. Normal unbulleted text has correct spacing next to table. Normal unbulleted text has correct spacing next to table.

---

Seems to work okay to me. Whaddaya think? DMahalko (talk) 00:02, 25 December 2009 (UTC)

Explanation

Per the CSS 2.1 spec, two types of boxes cannot overlap floats: a line box, and "a table, a block-level replaced element, or an element in the normal flow that establishes a new block formatting context (such as an element with 'overflow' other than 'visible')". The reason for this is pretty straightforward. CSS deals only with boxes. Imagine what would happen if no box could overlap a float. Then if you have even a short float, it will cause any adjoining block element to squish horizontally for its entire width, leaving an unsightly gap under the float:

Floated text

Some lengthy paragraph. Lorem ipsum dolor sit amet, consectetur adipisicing elit, sed do eiusmod tempor incididunt ut labore et dolore magna aliqua. Ut enim ad minim veniam, quis nostrud exercitation ullamco laboris nisi ut aliquip ex ea commodo consequat. Duis aute irure dolor in reprehenderit in voluptate velit esse cillum dolore eu fugiat nulla pariatur. Excepteur sint occaecat cupidatat non proident, sunt in culpa qui officia deserunt mollit anim id est laborum. Lorem ipsum dolor sit amet, consectetur adipisicing elit, sed do eiusmod tempor incididunt ut labore et dolore magna aliqua. Ut enim ad minim veniam, quis nostrud exercitation ullamco laboris nisi ut aliquip ex ea commodo consequat. Duis aute irure dolor in reprehenderit in voluptate velit esse cillum dolore eu fugiat nulla pariatur. Excepteur sint occaecat cupidatat non proident, sunt in culpa qui officia deserunt mollit anim id est laborum.

Instead, the block ignores the existence of the float, but each line is shrunk if necessary, so you get:

Floated text

Some lengthy paragraph. Lorem ipsum dolor sit amet, consectetur adipisicing elit, sed do eiusmod tempor incididunt ut labore et dolore magna aliqua. Ut enim ad minim veniam, quis nostrud exercitation ullamco laboris nisi ut aliquip ex ea commodo consequat. Duis aute irure dolor in reprehenderit in voluptate velit esse cillum dolore eu fugiat nulla pariatur. Excepteur sint occaecat cupidatat non proident, sunt in culpa qui officia deserunt mollit anim id est laborum. Lorem ipsum dolor sit amet, consectetur adipisicing elit, sed do eiusmod tempor incididunt ut labore et dolore magna aliqua. Ut enim ad minim veniam, quis nostrud exercitation ullamco laboris nisi ut aliquip ex ea commodo consequat. Duis aute irure dolor in reprehenderit in voluptate velit esse cillum dolore eu fugiat nulla pariatur. Excepteur sint occaecat cupidatat non proident, sunt in culpa qui officia deserunt mollit anim id est laborum.

Pay close attention to where the blue border is. The enclosing block is really behind the float, but the text doesn't go there. Thus we get a much closer approximation to what we want.

Unfortunately, this causes a problem: the padding on the inside of the blue-bordered float is way to the left of the left edge of the line boxes. Usually this isn't noticeable, because you put a margin on the floating element:

Floated text

Some lengthy paragraph. Lorem ipsum dolor sit amet, consectetur adipisicing elit, sed do eiusmod tempor incididunt ut labore et dolore magna aliqua. Ut enim ad minim veniam, quis nostrud exercitation ullamco laboris nisi ut aliquip ex ea commodo consequat. Duis aute irure dolor in reprehenderit in voluptate velit esse cillum dolore eu fugiat nulla pariatur. Excepteur sint occaecat cupidatat non proident, sunt in culpa qui officia deserunt mollit anim id est laborum. Lorem ipsum dolor sit amet, consectetur adipisicing elit, sed do eiusmod tempor incididunt ut labore et dolore magna aliqua. Ut enim ad minim veniam, quis nostrud exercitation ullamco laboris nisi ut aliquip ex ea commodo consequat. Duis aute irure dolor in reprehenderit in voluptate velit esse cillum dolore eu fugiat nulla pariatur. Excepteur sint occaecat cupidatat non proident, sunt in culpa qui officia deserunt mollit anim id est laborum.

The margin on the float forms part of the floated box, so the line boxes shrink to avoid it. The space on the left edge of the running text is caused by the float's margin where the float is, and by non-floated box's padding elsewhere.

Unfortunately, this causes problems in some hard-to-fix corner cases. We expect to see extra-large left padding on lists, but the margin you give to the float is fixed independent of whether a list is actually present. So if we have a list, it looks like:

Floated text

• Some lengthy paragraph. Lorem ipsum dolor sit amet, consectetur adipisicing elit, sed do eiusmod tempor incididunt ut labore et dolore magna aliqua. Ut enim ad minim veniam, quis nostrud exercitation ullamco laboris nisi ut aliquip ex ea commodo consequat. Duis aute irure dolor in reprehenderit in voluptate velit esse cillum dolore eu fugiat nulla pariatur. Excepteur sint occaecat cupidatat non proident, sunt in culpa qui officia deserunt mollit anim id est laborum. Lorem ipsum dolor sit amet, consectetur adipisicing elit, sed do eiusmod tempor incididunt ut labore et dolore magna aliqua. Ut enim ad minim veniam, quis nostrud exercitation ullamco laboris nisi ut aliquip ex ea commodo consequat. Duis aute irure dolor in reprehenderit in voluptate velit esse cillum dolore eu fugiat nulla pariatur. Excepteur sint occaecat cupidatat non proident, sunt in culpa qui officia deserunt mollit anim id est laborum.

Notice that the left padding is larger where the list's actual padding is used, but not where the float is. In this case you'd want to increase the right margin on the float:

Floated text

• Some lengthy paragraph. Lorem ipsum dolor sit amet, consectetur adipisicing elit, sed do eiusmod tempor incididunt ut labore et dolore magna aliqua. Ut enim ad minim veniam, quis nostrud exercitation ullamco laboris nisi ut aliquip ex ea commodo consequat. Duis aute irure dolor in reprehenderit in voluptate velit esse cillum dolore eu fugiat nulla pariatur. Excepteur sint occaecat cupidatat non proident, sunt in culpa qui officia deserunt mollit anim id est laborum. Lorem ipsum dolor sit amet, consectetur adipisicing elit, sed do eiusmod tempor incididunt ut labore et dolore magna aliqua. Ut enim ad minim veniam, quis nostrud exercitation ullamco laboris nisi ut aliquip ex ea commodo consequat. Duis aute irure dolor in reprehenderit in voluptate velit esse cillum dolore eu fugiat nulla pariatur. Excepteur sint occaecat cupidatat non proident, sunt in culpa qui officia deserunt mollit anim id est laborum.

IMO, this is the correct solution in our case. It's certainly much better than using table tags – or, as I did here, overflow:hidden, which achieves the same effect by creating a new block formatting context. The table-y solution gives the ugly blank space under the float I pointed out in my first example, so I'd strongly recommend it not be used. If you want to make manual tweaks so it looks better, increase the right margin on the float until it looks good (I used 1.5em here).

There's no obvious way to solve this on the spec level (at least, not obvious to me). What we would basically like is for the outer box to become no longer a box, like (simulated):

Floated text

Some lengthy paragraph. Lorem ipsum dolor sit amet, consectetur adipisicing elit, sed do eiusmod tempor incididunt ut labore et dolore magna aliqua. Ut enim ad minim veniam, quis nostrud exercitation ullamco laboris nisi ut aliquip ex ea commodo consequat. Duis aute irure dolor in reprehenderit in voluptate velit esse cillum dolore eu fugiat nulla pariatur. Excepteur sint occaecat cupidatat non proident, sunt in culpa qui officia deserunt mollit anim id est laborum. Lorem ipsum dolor sit amet, consectetur adipisicing elit, sed do eiusmod tempor incididunt ut labore et dolore magna aliqua. Ut enim ad minim veniam, quis nostrud exercitation ullamco laboris nisi ut aliquip ex ea commodo consequat. Duis aute irure dolor in reprehenderit in voluptate velit esse cillum dolore eu fugiat nulla pariatur. Excepteur sint occaecat cupidatat non proident, sunt in culpa qui officia deserunt mollit anim id est laborum.

But then the large blue-bordered region is no longer a box. CSS deals only with boxes; introducing anything that's not a box would cause enormous conceptual complications both for authors and implementers. So we have to sort of fake it. It works pretty well except for corner cases like this. Browsers are not going to "fix" this: it's not a bug in their implementations, and changing how this model works on any fundamental level would break sites all over the place. If any more convenient workaround is ever introduced, it would likely be tied to a new CSS property to enable the new behavior, whatever that might be (I can't think what).

To reiterate for those who didn't read all that: increase the right margin on the floated box, don't try using tables or similar. (And if you did want to use tables, overflow:hidden on each list item would probably be easier.) —Simetrical (talk • contribs) 19:52, 25 December 2009 (UTC)

Padding doesn't work with non-list text

Floated text
Floated text
Floated text
Floated text
Floated text

This unbulleted text is left-justified wrong, compared to the list content.

• One-line list example
• One-line list example

This unbulleted text is left-justified wrong, compared to the list content.
Another unbulleted line.
Another!
Foo.
Bar.
Baz.

---

The problem with adding padding as you are suggesting is that it also pushes over non-list content in a way that still looks wrong. Yes, you are giving more room for bullets and numbering, but the margin position is still wrong compared to non-list text.

To me it appears that enclosing each list item in its own non-float table is still the least-ugly way to deal with the formatting issues, until the SSL spec can be changed to fix this annoyance. DMahalko (talk) 05:21, 26 December 2009 (UTC)

Hmm, you're right. Using overflow:hidden to create a new block formatting context seems like a better tactic for short list items, while for very long list items you'd want to change the float's margin. There's no good solution here, though. —Simetrical (talk • contribs) 23:54, 26 December 2009 (UTC)

Question about this title: Oahspe: A New Bible

I use Watchlist sorter which sorts by namespace, and this article seems to have a namespace all of its own. Any ideas why? The colon in the title? Thanks. Dougweller (talk) 12:35, 26 December 2009 (UTC)

Yes, it's the colon. The source code of watchlist sorter even has the following comment: //This will fail for articles which contain ":" in name. Svick (talk) 14:27, 26 December 2009 (UTC)

Thanks. Dougweller (talk) 06:39, 27 December 2009 (UTC)

Javascript link in personal appeal header

Fail. —Preceding unsigned comment added by 92.28.188.226 (talk) 17:09, 26 December 2009 (UTC)

Template parameter default value

I created a new template here:http://en.wikipedia.org/w/index.php?title=Template:Hstrong's&action=edit. the code looks like this:
[http://www.htmlbible.com/sacrednamebiblecom/kjvstrongs/STRHEB{{{1|0}}}.htm#S{{{1}}}{{{2}}} {{{3}}}]
and it works fine 99% of the time. but to be perfect the first parameter (which i use twice) must, in the first instance, have a default value of 0 and in the second instance it must have a default value of null. but when i run {{Hstrong's||10|result}} the result is 'result' where the expected result would be 'result'. (note the 0 after the STRHEB)
Lemmiwinks2 (talk) 00:13, 27 December 2009 (UTC)

I have fixed your template. And here's the explanation:

Template parameters really have three states: undefined (not fed at all), empty but defined (fed but with no content, as in an empty string ""), and "has data".

When you feed {{Hstrong's||10|result}} then you are feeding an empty but defined parameter 1. But the parameter pipe trick you used for that parameter "{{{1|0}}}" only returns the default value 0 when parameter 1 is undefined (when it hasn't been fed at all). To make it return 0 for both an empty but defined, or a not defined parameter 1, do like this:

{{#if: {{{1|}}}    <!--Check if parameter 1 has data-->
| {{{1}}}    <!--Parameter 1 has data, use it-->
| 0    <!--Parameter 1 is empty or undefined, use a default value-->
}}

Note that you must use "{{#if:{{{1|}}}", not "{{#if:{{{1}}}". That is, there must be a pipe "|" in the if-usage, otherwise the if-case sees the literal string "{{{1}}}" instead of an empty string "" when parameter 1 is undefined.

For the second usage of parameter 1 in your code, the "#S{{{1}}}{{{2}}}" part, I assume you mean you want to simply get "#S" when parameter 1 and 2 are empty or undefined. Then do like this: "#S{{{1|}}}{{{2|}}}". That means when they are undefined they fall back to an empty string, and when they are defined but empty they instead return their content which happens to also be an empty string.

So we are working with empty and undefined stuff here, its pretty tricky.

--David Göthberg (talk) 01:52, 27 December 2009 (UTC)

Yes. that fixes it. thank you. :-) Lemmiwinks2 (talk) 02:42, 27 December 2009 (UTC)

I think in this case it might be easier to work with named parameters rather than positional parameters - something like:

[http://www.htmlbible.com/sacrednamebiblecom/kjvstrongs/STRHEB{{{page|0}}}.htm#S{{{page|0}}}{{{digit1|0}}}{{digit2|0}}}]

with named parameters there's less confusion - you'd have to feed it a blank entry specifically to get a blank. incidentally, is there a reason why you're entering the last two digits as independent variables rather than as one variable? --Ludwigs2 02:21, 27 December 2009 (UTC)

the last parameter is the text to show in the link. the first parameter is the page and the second parameter is the number within the page. page must default to 0 in the first instance only. in the second instance it must default to null. Lemmiwinks2 (talk) 02:33, 27 December 2009 (UTC)

ah. then just delete the zero in the second instance (chagned digit1 &digit 2 to more sensible names, as well):

[http://www.htmlbible.com/sacrednamebiblecom/kjvstrongs/STRHEB{{{page|0}}}.htm#S{{{page|}}}{{{number}}}{{text}}}]

Scottish island infobox map

Several people have raised the issue of maps such as on Barra and Rùm not appearing in IE8. Discussion. Can anyone cast some light on the problem? Finavon (talk) 10:57, 26 December 2009 (UTC)

For me, the maps display in normal mode in IE8 but usually not in compatibility mode. If I drag the mouse over the blank space in compatibility mode then the space where the map should be becomes completely blue and a blue arrow pointing up-right in a box appears outside the space. Placing the mouse on the arrow (no click needed) causes the map to display. PrimeHunter (talk) 12:09, 26 December 2009 (UTC)

The maps generally don't display for me, but I don't see any IE "Compatibility Icon" (the "torn page" icon) displayed for these pages. I'm assuming that this means I'm viewing them in "normal" mode? 81.129.128.20 (talk) 13:54, 26 December 2009 (UTC).

I always see the torn page icon to the right of the address bar on any page when I run IE8 whether compatibility mode is activated or not. The option is also in the page menu. There I see the torn page icon when it is in "normal" mode and a checkmark when it is in compatibility mode. PrimeHunter (talk) 17:15, 26 December 2009 (UTC)

Oh. According to the IE help, "If Internet Explorer recognizes a webpage that is not compatible, you will see the Compatibility View button ["torn page" icon] on the Address bar." I see this sporadically for some sites, but I don't recall seeing it for any Wikipedia pages, and I certainly don't get it for the problem Scottish island pages. I assumed this was because IE did not recognise any incompatibilities in those pages. I can force Compatibility View for the Wikipedia site via the menu option, but it does not appear to make any difference to the display (or in my case non-display) of the maps. All it seems to do is cause IE to crash a lot. 86.133.242.124 (talk) 00:25, 27 December 2009 (UTC).

IE8 is not my normal browser but after testing it on many pages I found a couple where the torn page icon was not at the address bar. Have you tried the mouse dragging thing? Hold down left click and drag the mouse to the right on the space where the map should be displayed. PrimeHunter (talk) 00:55, 27 December 2009 (UTC)

Yes, that works for me as you described earlier. When I hover the mouse over the blue "accelerators" arrow after having selected the blank space, the map appears. 86.133.242.124 (talk) 01:14, 27 December 2009 (UTC).

PrimeHunter, Finavon.. have you guys tried logging out ? Might be that there is a difference in IP vs. registered behaviour. (/me suspects the relative bodyContent that is applied when a sitenotice is active) —TheDJ (talk • contribs) 11:30, 27 December 2009 (UTC)

Return/get first word of page title

Would it be possible through the use of a template or any other means to be able to quote the certain words from a page's title? Consider this page's title for example "Village pump (technical)". Is it possible by some means to write a code so that the word "Village" is written/returned? (or if this 'code' or whatever is placed on some other page it writes/returns the first word from the title) Thanks.--72.241.13.250 (talk) 05:31, 27 December 2009 (UTC)

You can use the template {{first word}} I just created. Svick (talk) 10:38, 27 December 2009 (UTC)

Thanks so much! --72.241.13.250 (talk) 16:40, 27 December 2009 (UTC)

Can we change spacing/font in the box in which we edit articles?

I don't know about anyone else, but I find it frustratingly difficult to try to edit articles in the actual box in which the actual editing takes place... the font is small, it is single spaced and, most annoying, the html-type tags and references make it hard to tell what is actually going to appear in the article. A lot of references are really long so when you try to read paragraphs in the editing box its difficult to tell where one sentence ends and another begins independent of things inside of reference tags... I'm not sure I'm explaining this adequately, or making any sense... but I'm sure other people are annoyed by this!

When I try to edit a long article with lots of grammatical problems I like to just read through it in the actual box where I can make edits, but all the html tags and references clutter everything up so much that it is just very difficult to read...

Is there some way to improve this? For example, why can't it be more like editing posts in some forums on other websites where we could actually highlight text and italicize it (or bold it, etc) so that it shows up as italicized in the edit box rather than having to use apostrophes? If regular internet forums allow for this why can't Wikipedia do it? Is there anyway to improve the system for adding citations so that we don't have to put the entire bibliographic reference in the paragraph with the html tag around it? Could we make edit boxes double spaces so they are easier to read? Any ideas? The Way (talk) 06:53, 27 December 2009 (UTC)

You can use list-defined references to define references inside <references> (or {{reflist}}) and only point to them in the text.

You can change the appearance of Wikipedia for yourself by editing your monobook.css page (or vector.css if you use the beta skin). For example, to make the font of edit box larger, put #wpTextbox1 { font-size: 125% } there. Svick (talk) 11:00, 27 December 2009 (UTC)

As an addendum to Svick's suggestion, there's also the option to change the line-spacing. Separating each CSS declaration with semicolons (e.g. {foo:bar; foo2:bar2;}), add line-height:125%; or some such to increase spacing between lines. You can also change the font used; I added a Gadget a while ago that you can enable in your preferences to set the font to use a sans-serif font instead of the monospace that most browsers use. (Safari users get that by default, though.) {{Nihiltres|talk|edits}} 17:58, 27 December 2009 (UTC)

Ages ago I remember seeing a prototype of a new-style edit window where references, comments, HTML tags etc were all colour-coded. It looked pretty neat. Whatever happened to that, I wonder? It's a shame that it was never rolled out. 86.150.102.21 (talk) 20:09, 27 December 2009 (UTC).

If you're registered and not using Internet Explorer, you can enable the wikEd user script in the Gadgets section of your preferences. Some other syntax-highlighting (and template-folding!) work is likely to be rolled out as default through the Usability Initiative. {{Nihiltres|talk|edits}} 02:34, 28 December 2009 (UTC)

Orange links

I forget now where I saw it, but there was a guide to the different colors of links of Wikipedia, and orange wasn't on it. I have seen this color in at least two situations, although there might not be consistency in how the color change happens.

On one occasion today, a link was orange after I had clicked on it, though usually it's purple in those situations. Also, when making major improvements to an article or creating a new one, I find that when previewing, as I check each link to make sure it works properly, the next link for me to check turns orange.Vchimpanzee · talk · contributions · 22:32, 24 December 2009 (UTC)

That's nothing to do with Wikipedia, it's your Internet browser. If you click on a link, move the mouse away, and un-click, the link will be orange. ╟─TreasuryTag►constabulary─╢ 22:33, 24 December 2009 (UTC)

Even with article improvement?Vchimpanzee · talk · contributions · 22:52, 24 December 2009 (UTC)

I don't understand what you mean, "even with article improvement?" – if you click and drag away from any link, on any website (Wikipedia, Google, Department for Energy and Climate Change etc.), it will go orange. ╟─TreasuryTag►quaestor─╢ 22:55, 24 December 2009 (UTC)

The article (still being previewed) has new links which I have never tried. So after I try one, the next link for me to try, going left to right, is orange.Vchimpanzee · talk · contributions · 23:00, 24 December 2009 (UTC)

I guess it is the browser. While going through a cycle of clicking on links and then going back where I was, I see an orange link for where I just was.Vchimpanzee · talk · contributions · 18:06, 25 December 2009 (UTC)

It's the browser. Orange is the color for an "active" link (ie: when you're actually clicking it). EVula // talk // ☯ // 18:12, 25 December 2009 (UTC)

What EVula said. Most people don't realize that this happens because they finish clicking. :) --Izno (talk) 18:23, 25 December 2009 (UTC)

I think youwere looking for Wikipedia:Link color. ---— Gadget850 (Ed) ^talk 18:47, 26 December 2009 (UTC)

I didn't find my answer there, though I was thinking I had in the past. That's why I came here.Vchimpanzee · talk · contributions · 17:21, 28 December 2009 (UTC)

How to create a shortened version of a template that uses a table

I would like to create an option for a shortened version of a template that uses a table; the example is here and here. Unfortunately, I have not been able to get this process to work with templates that have tables, where we want to display only part of the table content in the shortened version. The template that this is intended for is one that is regularly updated, which is why it is preferable to use a shortened version of one template, rather than to create two separate templates. Any help with figuring this out would be greatly appreciated!--Pharos (talk) 02:14, 27 December 2009 (UTC)

I updated your sandbox with a table that does what you want. Here's the code for anyone else that reads this later:

<table class="wikitable">
<tr>
<td> Cell A
<td> Cell B
{{#ifeq: {{lc: {{{short|}}} }} | yes
| <!--Short version, do nothing-->
| <!--Long version-->
  <tr>
  <td> Cell C
  <td> Cell D
}}
</table>

It uses HTML table notation instead of wikitable notation, so we don't get pipe "|" problems inside the #ifeq. It is usually not necessary to add ending </td> and </tr> tags, since MediaWiki fixes that for us. But if you take cell content as parameters, then you might need ending </td> tags. Then do like this: "<td> {{{1|}}} </td>", not this: "<td>{{{1|}}}</td>", to avoid some tricky whitespace problems. Do not put the ending </td> tags on a separate line, that also causes sneaky whitespace and padding problems.

"{{lc: }}" lower cases the "short" parameter, in case people do the mistake to feed "short=Yes" or "short=YES".

--David Göthberg (talk) 03:11, 27 December 2009 (UTC)

You should not make assumptions about what the mediawiki software will or will not fix "for you". It can change and often without notification. —TheDJ (talk • contribs) 11:27, 27 December 2009 (UTC)

The reason I often leave out the end <td> and <tr> tags in my examples is that humans tend to place them wrongly and mess them up in any number of ways. Among other things as I described above, even if you place them where most users would expect to be "correct" you get whitespace and padding problems. While when you let MediaWiki handle it you usually don't get those whitespace and padding problems. So MediaWiki does a better job than the human editors.

--David Göthberg (talk) 11:53, 27 December 2009 (UTC)

These are probably being fixed by Tidy, not part of the MediaWiki parser. When people try to copy Wikipedia content to other sites, it breaks horribly because they don't have Tidy installed and the HTML in the template is completely broken. Tidy should be used to fix mistakes, not to take shortcuts. Things like this also make it difficult to add to or fix templates, because you have to work around the fact that they're intentionally "broken." Mr.Z-man 17:08, 27 December 2009 (UTC)

Isn't that example valid HTML5? Happy‑melon 11:55, 28 December 2009 (UTC)

4.01, at the least (we're running xhtml 1.0 transitional, of course). It's only valid html5 if MediaWiki plans to move to the html serialization rather than the xhtml serialization of html5. --Izno (talk) 20:18, 28 December 2009 (UTC)

My reading of the spec doesn't say <tr> and <td> tags can ever, ever be left unclosed. Can you point me to where you got this idea? --MZMcBride (talk) 00:50, 29 December 2009 (UTC)

This is normal behavior for html 4 (and earlier) and the html serialization of html5. XHTML by it's nature as an XML document must always have its tags closed. --Izno (talk) 01:32, 29 December 2009 (UTC)

Mr.Z-man: We can't take into account what set-ups, what versions of MediaWiki, what CSS in Common.css, and what meta-templates other wikis have installed. If we bothered about that then most of the template coding we do here would not be possible or be horribly bloated with extra workarounds and fixes. The templates we code here are intended to be used here, not somewhere else.

The only thing we can do is to document our templates properly, including describing their technical details, so those that want to reuse them elsewhere have a fair chance to understand how they work.

--David Göthberg (talk) 00:45, 29 December 2009 (UTC)

In order to make Wikipedia as useful as possible, it has to be possible to reproduce the content elsewhere. I agree with you, David, about documenting quirks well. But I also agree with Mr.Z-man that we should (ideally) only ever rely on Tidy to catch mistakes. --MZMcBride (talk) 00:50, 29 December 2009 (UTC)

CSS can easily be copied, as can meta-templates if using Special:Export, and there are very few backward incompatibilities in MediaWiki. Templates that purposely use broken syntax are much harder to fix. The templates we use here are designed for use in our content, which we explicitly encourage copying of. Mr.Z-man 02:47, 29 December 2009 (UTC)

Filter acting against content guideline

Hello, I recently tried to blank User talk:86.44.24.15 after that IP was assigned to me, to clear an old warning. I was not allowed to make this edit, even though it is per WP:BLANKING. It is bad enough when innocent edits are given the mark of Cain by some edit summary tag, but to be disallowed from making an edit? If you do not want IP edits then go ahead and say so. 86.44.55.218 (talk) 11:37, 28 December 2009 (UTC)

You must have a dynamic IP address which changed before you tried to blank. The filter logs show no attempt to blank User talk:86.44.24.15 by that IP [4] but 4 attempts by the IP 86.44.55.218 [5] you posted with here. All 4 were correctly stopped with the explanation "Unregistered user blanking someone else's user talk page". PrimeHunter (talk) 12:18, 28 December 2009 (UTC)

watchlist rss feed - no section links?

I've been experimenting with the watchlist rss feed, and fouind it really clunky to use. the most glaring deficit, of course, is that links in the feed are always generic links to the top of the page in question - no section links, no links for diffs or editing, etc. I can't imagine that the first (at least) would be difficult to achieve, so either that's an oversight on the server end or a misconfiguration on my end. is there a way to get that? --Ludwigs2 16:47, 28 December 2009 (UTC)

"Count page watchers"

Is this really what it sounds like, why is it allowed if the facility is not allowed on-wiki, and the same issues go for the opening-up of Special:Unwatchedpages. ╟─TreasuryTag►cabinet─╢ 22:15, 28 December 2009 (UTC)

I guess the easiest comparison is that Toolserver users have access to aggregate stats about deleted edits and most users don't get access to that on-wiki as well. watcher.py restricts the count if the page has fewer than 30 watchers, though there's an access list at Meta to override that. It might be something that would be nice to implement for admins into core MediaWiki; not sure where you'd stick it in the interface, though. --MZMcBride (talk) 22:23, 28 December 2009 (UTC)

Will this bug ever be fixed?

Hi

Could someone with some influence PLEASE, PLEASE try to get the stupid bugs in Wikipedia's diff generation fixed. Here is a typical example:

http://en.wikipedia.org/w/index.php?title=Goodwin_Sands&action=historysubmit&diff=334191631&oldid=334190790

There are only minor changes to odd words throughout this section, but the code is obviously getting confused at the outset because an extra line has been inserted at the start. Then it flags virtually the whole section as having changed.

This bug has been outstanding for YEARS. It is so dumb, and so annoying, it drives me nuts. Would someone PLEASE try to get it fixed. Thank you! 86.133.242.124 (talk) 02:51, 27 December 2009 (UTC).

ha! See below. Similar bug. ~ R.T.G 04:19, 27 December 2009 (UTC)

I've seen this plenty of times, too. It is very misleading. I submitted a bug report on this: bugzilla:21953. Please vote on it if you want it noticed. stmrlbs|talk 05:05, 27 December 2009 (UTC)

I'm sorry, the diff I gave below is a good example of a bug, significant new text not highlighted in the diff, but I won't be posting my email on a public page again. I have had useless stupid flooding on multiple emails (receive 100s unsolicited emails a day!). I had an email with my own name on it but [email protected] decided to inform me of unrequited rubbish for ever. I have an SUL. My username idents me as good as my email and you can email me on the hush without ever seeing my email@add. If they fix it so that I can log in without displaying email I will vote and post this diff. Otherwise I won't. ~ R.T.G 11:59, 27 December 2009 (UTC)

Create a throwaway/single-purpose Yahoo account or the like.--Father Goose (talk) 12:48, 27 December 2009 (UTC)

Yeah but logging in privately would be good wouldn't it... it's not the first time I have decided not to participate something which wanted fake emails. Sorry. Still would be good to see this one fixed though. ~ R.T.G 20:17, 27 December 2009 (UTC)

Diffs

I need to see this[6] as a diff rather than an undefined red paragraph and a black one, which on examination, has significant changes in the middle. Current highlighting does not even almost highlight a diff in some edits at all and yet there are new words and links and etc.. Obviously that is probably a core program area. Nonetheless - that diff is an example of diff showing no diff although diff.. If I wanted to I could change a large 200 word paragraph and the diff would show nothing as if the paragraph had been moved only rather than changed in the actual diff. I can find the !significant! difference between those paragraphs (see diff) but my computer and wiki does not which is, after all, defeating the point of diffs in the strongest way possible. Diffs without diffs tells us no diffs and yet... hmmm... please show diffs if they are existing rather than red paragraphs, for any truly able tech help, thank you. Without... thank you anyway but should report anything possible for a fix, probably old post but, should repost. ~ R.T.G 04:16, 27 December 2009 (UTC)

Diffs are incredibly difficult to do well in all cases. What we have at the moment is basically the best you will get. If you think you can do better, please do so, submit your code to the devs and maybe publish a couple of papers. If not, then either accept what we have or go do a PhD on the subject. In the mean time, try not to change the number of blank lines between paragraphs while you edit and the resulting diffs should be much clearer. OrangeDog (τ • ε) 20:30, 27 December 2009 (UTC)

I don't really understand the terms under which this kind of Wikipedia software development takes place, but by any normal criteria this response is, frankly, not good enough. This is not an obscure or exotic bug. It is a basic and obvious failure of the code to do something that it quite obviously should be capable of doing, such as recognise that a line has been inserted into an otherwise unchanged section. If the developers can't fix this, then more capable ones need to be recruited. 86.150.102.21 (talk) 22:44, 27 December 2009 (UTC).

There are few developers and they have a lot to do. See User:Cacycle/wikEdDiff for an alternative diff which often gives better results. PrimeHunter (talk) 23:52, 27 December 2009 (UTC)

Insufficient resources is a different matter. That is not what OrangeDog said. OrangeDog implied that the problem would not be fixed because it is too difficult to fix, which I do not believe is a satisfactory answer. 86.150.102.21 (talk) 00:51, 28 December 2009 (UTC).

OrangeDog is correct. Diff generation is very difficult to do well, and even more difficult to do it well and efficiently. The diff algorithm needs to be able to handle huge diffs on large articles without killing the servers, but it also needs to be fine-grained enough to handle small changes. However, killing the servers is a much bigger problem than not being quite good enough, so it tends to get the priority. Mr.Z-man 00:16, 28 December 2009 (UTC)

We're not asking that it should work to an unreasonable level of perfection, or perform some fabulously difficult task. Even the most basic diff generator should be able to figure out that a blank line has been inserted. It's not rocket science. 86.150.102.21 (talk) 00:56, 28 December 2009 (UTC)

You say that with such confidence, but that is precisely the problem. If it were "rocket science" (that is, essentially just complicated mathematics) then the software would find it much easier to handle. Trying to get a computer to solve an abstract problem, like diff generation, is infinitely harder; it is a "fabulously difficult task". You blindly assert that "even the most basic diff generator should be able to [do X]", but that's simply not true, because 'doing X' is actually a seriously complicated problem for a computer to solve, and solve efficiently. Happy‑melon 11:52, 28 December 2009 (UTC)

I used to earn a living as a computer programmer, so I am not completely unaware of what I am talking about. I'm sure it is indeed hard to write an efficient diff generator that works "perfectly" in all circumstances, but here we are not talking about "all circumstances", we are talking about detecting simple changes such as the insertion of a line at the start of a section. This is not an exotic "nice to have" feature. This is the nuts and bolts of what any reasonable diff generator should do. Wikipedia's failure to do it successfully is a careless bug -- simple as that. 81.129.130.232 (talk) 12:12, 28 December 2009 (UTC).

I know that there are different algorithms for comparing differences, and it isn't an easy thing to program. However, I don't think the point is how hard is it to check for a blank line, but rather that inserting a blank line is a rather sommon thing that editors do in Wikipedia. I think many editors will add a blank line after a heading, in between paragraphs and sections, and in between readers' comments in threads., because it makes the text easier to read. So, I think that whatever algorithm Wikipedia uses, it should take into account common editing patterns and try to incorporate checking for these common differences. stmrlbs|talk 00:41, 29 December 2009 (UTC)

just as an informal benchmark, I double-checked the above diffs using TextWrangler's compare feature (which I think uses the standard diff routine built into unix). it did a better job catching the differences on the 'Rockall' diff, but failed as badly on the 'Goodwin Sands' diff. is wikipedia's diff routine written in PHP, or is it a prefab, pre-compiled thingamabob? I'm curious about the process now, and I'd love to take a look at the source. --Ludwigs2 18:20, 28 December 2009 (UTC)

The documentation of classes that handle diffs in MediaWiki is at [7]. The pages of particular classes have links to their source code. Svick (talk) 19:01, 28 December 2009 (UTC)

Interestingly, looking at both of those diffs with popups gives better results. OrangeDog (τ • ε) 20:03, 28 December 2009 (UTC)

Popups does a word by word comparison, and only displays 1 paragraph of diffs. As such it is heavily flawed in other parts of it's algorithm. —TheDJ (talk • contribs) 14:40, 29 December 2009 (UTC)

Popups didn't make a big change to that diff I gave, sorry. edWiki doesn't work on Internet Explorer so can't say if that is any good. I don't think telling me to get a PhD in wikisource code and publish papers worked either! I don't want to self loathe myself for complaining about a gift which I really appreciate. Although the source area for the diff code appears extremely complicated, it is difficult to understand if there is not a basic part which checks character by character like WP:AWB does (don't forget, in the diff I provided a whole block of text where the diff was four or five consecutive words showed up that nothing except a space had been changed, the diffs further up show something more complicated which I have not examined, many singular words throughout a large diff). I thought there would be little or no activity on programming these days. I think the interface is fine, I mean fine, especilly in comparison to the attempts to change it on Wikia which was altered into something crap the last time I tried (no offence but it was truly difficult to edit while on this interface editing is only difficult when finding your place among refs, I do not doubt the changes in Wikia were a good or progressing effort beyond what I am capable of) ~ R.T.G 16:49, 29 December 2009 (UTC)

Quote: "... the source area for the diff code appears extremely complicated, it is difficult to understand ..." I suspect that this is the real reason why this glaring bug has never been fixed, and shows no signs of ever being fixed. Every developer who looks at the code probably recoils in horror and quickly finds something else to do instead. 86.152.242.96 (talk) 22:24, 29 December 2009 (UTC).

Instructions for Cacycle WikEdDiff

The Javascript gadget provided by Cacycle seems to handle this. To use it,

1. go to your watchlist
2. click on your preferences (look at the upper far right corner in the standard layout to find preferences)
3. click Gadgets
4. check wikiEd (under Editing gadgets )
5. then go to edit to bring up the wikiEdit box
6. toggle "wikEdDiff" button on User:Cacycle/wikEd_help

from now on, you should be able to see a 2nd variation of the differences (just under the regular revision history change display). stmrlbs|talk 00:42, 28 December 2009 (UTC)

If anyone who does these things is reading, a blasphemous port to Internet explorer would be much appreciated. I do not know what that involves only pointing out that it's not een available so go on... forget your New Years and sit at home scratching out text now!! ~ R.T.G 16:53, 29 December 2009 (UTC)

Using strict transport security for login

I have chronic problems where wikipedia doesn't seem to maintain my session data. In one browser tab, I login to wikipedia using the Secure Login. When I open a new tab, and hit en.wikipedia.org (either by typing it directly, or through a google search), it claims that I'm "not logged in". This doesn't pass the "session management smell test", since amazon, google, yahoo, pretty much every other major web site on the Internet works in the expected way.

I want anything related to authentication to wikipedia to use a secure session. Although I understand that many may not feel the need to use it, I don't like the idea of my password or authentication tokens being transmitted in the clear. It is perfectly OK if I can set a preference to force secure login, but allow edits to be made in the clear. I see archived discussion about the load being too much, but that would only be the case if the entire edit session was using SSL, not just the login.

Is there any chance the wikimedia software will be updated at some point to support Strict Transport Security or Strict Transport Security for login and cookie transmission? Is there a consensus that this should happen? (That's separate from the consensus/agreement on how to implement it) --rhyre (talk) 01:21, 28 December 2009 (UTC)

I'm not sure I understand correctly, but keep in mind that http://en.wikipedia.org/ and https://secure.wikimedia.org/wikipedia/en/wiki are different sites. Logging in via secure login (the second webpage) will not log you in at the first. So, when a Google search sends you to http://en.wikipedia.org/wiki/Earth and you find out you're not logged in, it's because you're not logged in to en.wikipedia.org. You're logged into secure.wikimedia.org. — The Earwig @ 04:56, 28 December 2009 (UTC)

That sounds broken to me. Why can't a secure login log you in at en? I thought we had cross site login. In any case, the lack of secure login for everyone compromises the integrity of Wikipedia. At the least I'd like a user option to require secure login (it would kick me to the secure site if I tried to login insecurely) and I think that only secure login should be allowed for admins.--agr (talk) 12:07, 28 December 2009 (UTC)

Let's say you log in on secure.wikimedia.org, and then visited en.wikipedia.org. If cross-site login was implemented, your login cookie from the secure site would be sent to en.wikipedia.org over insecure HTTP, compromising your secure login. Zetawoof_(ζ) 07:35, 29 December 2009 (UTC)

Fix broken AfD

As usual, when I try to nom an AfD without an automated tool, I've made a right royal mess of it.

If anyone can fix Wikipedia:Articles for deletion/Coal oil lamp, I'd be most grateful.

Even more grateful if someone could work out what I did wrong and fix the AfD instructions to make them even more idiot-proof. --Dweller (talk) 09:22, 29 December 2009 (UTC)

Hey there. The template you fill out to get the discussion page is {{subst:afd2 | pg=PageName | cat=Category | text=Why the page should be deleted}} ~~~~. To create the page you did originally, you must have set the page name ( pg= ) section to {{subst:Coal oil lamp}} - if you test this as a preview in a sandbox, you will find it gives you exactly the same weird screen that you got. How did you first create the discussion page? Ale_Jrb^talk 11:38, 29 December 2009 (UTC)

Thanks. I clicked "preloaded debate" from the AfD box. I guess I may have pasted in the subst code instead of the page name. In which case the instructions are idiot-proof, it's just that some idiots can't follow instructions. Thanks --Dweller (talk) 11:52, 29 December 2009 (UTC)

Oh - I see it. When you click the preloaded debate button, it creates the code {{subst:afd2|pg={{subst:SUBPAGENAME}}|cat=U|text=Reason}} ~~~~ on the AfD page. Because the page is automatically created in the right place, when you save it the {{subst:SUBPAGENAME}} will be replaced with the name of the article you are nominating. You must have changed the SUBPAGENAME part to the name, which caused the problem. For reference, there is no need to do this. Ale_Jrb^talk 12:03, 29 December 2009 (UTC)

Audio player display weirdness

Hi. Yesterday I noticed that the audio player at the bottom of the infobox in While My Guitar Gently Weeps was oddly off to the right. I've managed to narrow the issue down to a minimum test case (see User:guillom/test). If someone could take a look at it and identify the actual issue, it would be much appreciated. The template to be fixed is {{Extra musicsample}}. Many thanks. guillom 18:18, 29 December 2009 (UTC)

The player is centered under the songtitle. —TheDJ (talk • contribs) 19:05, 29 December 2009 (UTC)

Unless you happen to have mwEmbed Gadget enabled in your preferences. In that case, some minor inconsistencies are expected in this case, due to the differing width of the old and the new player. —TheDJ (talk • contribs) 19:14, 29 December 2009 (UTC)

Yes, I have the mwEmbed gadget enabled. Could you please point me to the exact source of the issue, so that we make sure we fix this before releasing the player in production? guillom 19:34, 29 December 2009 (UTC)

The source of the problem is that both players have different default widths. When the switch occurs, we can just adapt {{Extra musicsample}}. —TheDJ (talk • contribs) 00:46, 30 December 2009 (UTC)

I fixed this. {{Extra musicsample}} was using a trick to center the player, but that trick was unnecessary and broken. —TheDJ (talk • contribs) 01:43, 30 December 2009 (UTC)

Show/hide in tables

Is it possible to change the show/hide toggle text in collapsible tables? I mean, from [show] to something like [view], or perhaps even [view | edit]. Intelligentsium 00:24, 30 December 2009 (UTC)

Yes, it's possible - the code is in Mediawiki:Common.js: var collapseCaption = "hide"; var expandCaption = "show"; - of course, changing this would change it for everyone, so it's unlikely to happen. As far as I'm aware there is no way of changing it for an individual table, short or changing everything or using your own copy of the Javascript. Ale_Jrb^talk 00:27, 30 December 2009 (UTC)

Fast switch between monobook and vector skin

Is it possible to switch between monobook and vector skin by just one or two mouseclicks? The way via “Try Beta” or "My preferences" is too long. --Leyo 09:25, 22 December 2009 (UTC)

You can change your skin temporarily by appending ?useskin=vector or ?useskin=monobook to the URL, but as soon as you click a link, you have your old skin back. Svick (talk) 12:04, 22 December 2009 (UTC)

You could write a script that gives you a button or link to do this. OrangeDog (τ • ε) 13:13, 22 December 2009 (UTC)

Thanks for the replies. I have been aware of the option suggested by Svick, but I would like to keep one skin till I click again. I was actually thinking of a script, but unfortunately I am not able to write such a script myself.

The main goal of the fast skin switch is to quickly enable/disable scripts/gadgets imported to these skins. --Leyo 16:28, 22 December 2009 (UTC)

Why switch skins for that? Better use a control that sets/unsets a cookie, and then you can conditionally load (sets of) scripts based on that. No need to switch a skin for that. —TheDJ (talk • contribs) 19:59, 22 December 2009 (UTC)

I tried something similar (on my de-WP account), but without success. Unfortunately, I am a beginner in JS. --Leyo 08:38, 23 December 2009 (UTC)

You can use another user account for that. And then access that account from a second browser.

Or you can use the secure server to access the second account. That lets you use different tools in different tabs in the same browser at the same time, since you can then run two different user accounts at the same time from the same browser.

To set that up do this: Open a new tab in your browser, go to Special:UserLogin, choose the secure server, create a new user account say "Leyo2". Then add whatever tools you want to that user account. Then when you use Wikipedia you can log in as Leyo to the normal servers, and as Leyo2 to the secure server. Nifty, isn't it? Oh, and don't forget to add some texts on the user page of Leyo2 that it is a secondary user account of Leyo, and redirect Leyo2's talk page to your normal talk page.

--David Göthberg (talk) 01:02, 24 December 2009 (UTC)

Thank you for this interesting suggestion (explained in detail). I would however like to use one account only.

I would like to know more about the proposal by TheDJ. I think this method could also be interesting for other users (as might be the proposal by David Göthberg, too). --Leyo 08:24, 24 December 2009 (UTC)

Yeah, TheDJ's suggestion to use a script to quickly load and unload the other scripts would be the most practical. However some javascript expert needs to code it up.

My method to use two accounts (in two browsers or with the secure server) is a bit clumsy and pretty annoying to use. By the way, my detailed description was mostly meant for the other readers of this page, since when we write a short explanation we usually get follow up questions, so I have gotten into the habit to write a full explanation from starters.

--David Göthberg (talk) 01:26, 26 December 2009 (UTC)

If I have time, I'll try to make such a thing this week. —TheDJ (talk • contribs) 14:04, 29 December 2009 (UTC)

That would be great, thanks. --Leyo 11:58, 30 December 2009 (UTC)

Is nobody else having problems with viewing history pages?

I reported this several days ago, and nobody made any mention of also having the problem, but I am having SEVERE problems with the loading of history pages. I have tried both Firefox and IE, and the problem exists in both browsers. The proper history page loads quickly, but then the page freezes for almost a minute, I can't do anything until the page unfreezes. If I attempt to do anything on the page, such as go Back to a previous page or attempt to view a diff, the browser logo (on both IE AND Firefox) in the extreme upper left corner of my screen changes from the Firefox fox or the IE "e" to a white box with blue borders, and I can't do a thing until it unfreezes. Woogee (talk) 23:00, 29 December 2009 (UTC)

I had a hell of a time with freezups for a while until I discovered that the current version of Firefox has terrible memory leaks in terms of its java/javascript handling. Try installing NoScript and also perform the other Firefox speed tweaks you can find by Googling "speed up firefox". –xeno^talk 23:02, 29 December 2009 (UTC)

I did both of those things, and it's still freezing on history pages. But thanks. Woogee (talk) 23:35, 29 December 2009 (UTC)

Yea, FF 3.5.6 (the current version) has a really bad memory leak (Again!). Sucks.
— V = I * R (talk to Ohms law) 23:55, 29 December 2009 (UTC)

Indeed, FF 3.5 in general has had trouble with the way it handles long pages with many form elements (our history pages being just about the canonical worst case for this). 3.5.0 was frankly unusable on history pages, and 3.5.6 seems to have a similar issue, though not as bad. Since we're already implementing the suggested workaround (give every form element a unique "id=foo" attribute), there's not much we can do on our end except wait for a new (competent) release. — Gavia immer (talk) 01:03, 30 December 2009 (UTC)

But I also have this freezing problem in IE. Woogee (talk) 06:18, 30 December 2009 (UTC)

Both browsers? That's strange. What versions of IE and Firefox are you using? On what operating system?

I see you have no subpages in your user space, so you don't seem to have any user scripts installed. Test if you get the same freezing of history pages when you are logged out.

If you don't get the freezing when logged out, then I suspect it might be some Wikipedia gadget you are using. See the menu at the top of the page - my preferences - Gadgets. And if turning off the gadgets worked, then please tell us which gadget was causing it so we can fix it.

If you do get the freezing when logged out: You said that both browser's logos are affected during the freeze. So check if you have any extra toolbars or other plug-ins installed in both of your browsers.

--David Göthberg (talk) 18:15, 30 December 2009 (UTC)

enhanced watch list abbreviations

(In Russian, see: ru:ВП:Ф-Т#комментарии в списке наблюдения)

Some time ago, enhanced watch list design (in Russian) was changed: marks like "б" (bot), "м" (minor), "Н" (new) was become as commented text. For example, мб:

<abbr class="minor" title="Эта правка помечена как малая">м</abbr><abbr class="bot" title="Эта правка сделана ботом">б</abbr>

Problem with this is that such texts not searched through Ctrl F in Opera (neither "мб" for combinations, nor "м" or "б" for separate letters); not mentioning, that this makes more traffic. How (who may/should do) to remove this change? A^VB^talk 02:11, 30 December 2009 (UTC)

English translation is also affected: with enhanced watch list we have in watch list "mb"

<abbr class="minor" title="This is a minor edit">m</abbr><abbr class="bot" title="This edit was performed by a bot">b</abbr>

and searching "mb" above by Ctrl F in Opera doesn't work. Don't know if other browsers affected or not, but weight of watch page these comments anyway increase. A^VB^talk 10:58, 30 December 2009 (UTC)

It doesn't significantly add to anything that we need to worry about. (ie: the servers aren't going to buckle under the added code, otherwise it would've been turned off; that change happened here several months ago) The text isn't searchable because browsers don't look at tooltip text. EVula // talk // ☯ // 11:24, 30 December 2009 (UTC)

He means that the combination mb is not searchable. This is caused by the abbr spans around the text. A minor issue I would say. —TheDJ (talk • contribs) 11:39, 30 December 2009 (UTC)

Ah, okay, misread that part. I agree that it isn't an especially big deal... especially since it apparently varies from browser to browser; I'm using Safari/Mac, and I could find both "мб" and "mb". If it's just an issue with Opera, well, that's not really our fault. :) EVula // talk // ☯ // 11:48, 30 December 2009 (UTC)

• caused by the abbr spans - interesting, that plain "span" doesn't affected, try to search this: mb

<span title="This is a minor edit" style="border-bottom:1px dotted; cursor:help; white-space:nowrap; font-weight:bold">m</span><span title="This edit was performed by a bot" style="border-bottom:1px dotted; cursor:help; white-space:nowrap; font-weight:bold">b</span>

minor issue - ...for those, who doesn't search in Opera... :( Is there reasons, why abbreviations can't be removed (as this was earlier)? A^VB^talk 16:10, 30 December 2009 (UTC)

I could find that "mb" as well. It really does appear to be an Opera-related issue. (and given Opera's 1.45% market share, I'd say the problem is solely on their end, and we don't need to do anything) However, take my position with a grain of salt; I'm just not a proponent of extensive browser workarounds for browser deficiencies (though Opera is far and away better than IE in that regard).
The abbreviations were added because they add a degree of functionality for those with non-traditional browsing devices (such as blind users that rely on screen readers). A simple "m" (that is only formatted visually) means nothing; a tag saying that it stands for a minor edit gives the character more meaning. Given the Foundation's goal of making everything as accessible as possible, the needs of low-vision users should be taken into consideration. EVula // talk // ☯ // 18:00, 30 December 2009 (UTC)

• find that "mb" as well - ...and this is what I say: plain span not affected. 1.45% market share - 2,31% worldwide; in Russia Opera measured from 16-24% to 36-38%. In Ukraine - 42-46%. Opera-related issue - page size increase isn't any browser dependent. taken into consideration - can't say for low-vision (I myself with broken vision, BTW), but broken search is affect me. If no one want to decrease page weight be removing abbreviations, why not use "span" instead "abbr"? A^VB^talk 23:14, 30 December 2009 (UTC)

The problem is that this is, basically, a correct use of the abbr tag. The tag is designed to provide further information and increase rather than decrease functionality by, like EVula says, to non-standard browser systems etc. (see) - and this benefit is not retained by switching to span tags. It's unfortunate that Opera doesn't handle it correctly, but that's more their problem than ours, and I'm not sure whether the overall benefit would be greater by keeping or changing it. Ale_Jrb^talk 23:24, 30 December 2009 (UTC)

Font problems

Hey guys. Have any changes been made to MediaWiki that would make the font appear differently to readers? I haven't noticed any change in font myself from multiple computers (with distinct browsers and OS's), but OTRS has received a couple of complaints in the last week that the font has changed, and I'd never seen a complain like that before... Just hoping someone knows what's going on. Someguy1221 (talk) 08:44, 30 December 2009 (UTC)

There have been font changes to tt, pre and code elements. diff. Can you pinpoint it to such elements ? I don't remember seeing any other changes, except that there is an awful big annoying banner at the top that might throw people off. —TheDJ (talk • contribs) 11:17, 30 December 2009 (UTC)

We have also updated the font sizes for the <source> tags, and for content shown on .css and .js pages. See this diff.

The discussion about the fixes to the <tt>, <code>, <pre>, and <source> tags is at MediaWiki talk:Common.css#Teletype style fix for Chrome. We did those fixes since some browsers have a bug that caused text with those tags to be very tiny.

Someguy1221: Could you tell more details from the messages you got through OTRS? And the best place to report that is to MediaWiki talk:Common.css.

--David Göthberg (talk) 17:58, 30 December 2009 (UTC)

Thanks all. I'll try to get more feedback from the readers who filed complaints, to try and pinpoint the problem. Someguy1221 (talk) 19:30, 30 December 2009 (UTC)

Rights changes

Why is it that in some users' rights log, when they gain the sysop bit, the log says something like:

• Time, day, month, year Some bureaucrat (Talk | contribs) changed group membership for User:some user from (none) to (none) ‎ ( sysop)?

Why does it say (none) to (none) when added the bit? Intelligentsium 20:40, 30 December 2009 (UTC)

Without seeing an actual log entry, my first reaction is that it's because the 'crat forgot to check the "admin" box. EVula // talk // ☯ // 21:12, 30 December 2009 (UTC)

Would that still create an entry? I remember the user did become an admin (but there were no additional entries), but was later desysopped. Could that have something to do with it? Intelligentsium 21:47, 30 December 2009 (UTC)

I believe so, yes. –xeno^talk 21:50, 30 December 2009 (UTC)

The old format of the log used to only say the right changed, for example, mine would have looked like this:

• 22:24, 3 April 2006 Essjay set rights for "User:Prodego" ( sysop)

Since then, the log format has been changed to show the rights the user had before, as well as the rights the user has after. The " sysop" text has been replaced with a customizable comment. Formerly there was no spot to comment when making rights changes. So now, all of these old entries have blank entries for the rights of the user before and after the change (since it wasn't stored then) and the comment contains what right was changed. Prodego ^talk 21:53, 30 December 2009 (UTC)

Why are some SVGs squashy?

In Taxation in the United Kingdom, see the two pie charts under "[["residence and domicile" and [["income tax". You can see that the first is thumbnailed neatly, the second clearly not (and the second also has an aspect ratio distortion, on my screen). Can a guru tell me why - as they are both SVGs, you can see their sources at File:UK taxes.svg (nice) and File:UKExpenditure.svg (nasty). I assume it is something to do with the viewbox (?) but I've deliberately manually treated both files in the same way in that respect, as you can see in the source. These were both edited with Inkscape under Windows. Thanks, Splash - tk 01:08, 31 December 2009 (UTC)

It's because Wikipedia image rendering has a problem with some fonts in SVG files. And it looks like you are using different fonts in the two files. Change all text items in File:UKExpenditure.svg to use the same font as you use for the text in File:UK taxes.svg, and it should look fine in the article.

After you uploaded the new version of the image you usually need to wait a minute for the servers to update, then bypass your browser cache so your browser loads the new version of the image, then purge the article page. (You don't always need to do all those steps.)

--David Göthberg (talk) 01:42, 31 December 2009 (UTC)

Image aspect problem

I would be most grateful if someone could explain me this glitch: please look at the front image of Optical_fiber#Termination_and_splicing featuring ends of two cables. Don't know what I see, but I see an aspect distorted image File:ST-optical-fiber-connector-hdr-0a.jpg. I have cropped it yesterday at commons and it still looks distorted on my PC now, though I tried to purge the page - in fact, I always see it this way if I change the aspect ratio (cropping with GIMP) of an image. Sorry for bothering - I am sure this is a well-known issue. Materialscientist (talk) 04:30, 31 December 2009 (UTC)

Sometimes the servers fail to update the cached copies of the image. One trick is then to set the image thumbnail to a new previously unused size, say 201px, thus forcing the image servers to render a fresh image for that size. I did so in the article for you. Unfortunately that means you can "never" again use the old size.

There also is a method to make the servers re-render (or drop?) the cached copies of the image, thus giving you correct rendering even in the previously used sizes. I looked around but couldn't find that method in the image help files. I know users here at the Village Pump have shown the method when this question has come up before. I'll see if I can dig it up and document it in the image help files.

And we really should add a button to the image pages that says "Purge this image" so we can easily perform that special purge, since as you noticed normal purge doesn't purge the cached copies of the image.

--David Göthberg (talk) 05:11, 31 December 2009 (UTC)

Thanks! Just a note that the full-scaled image is always fine - no need to purge there, it is the thumbed version (that is "|thumb" operator in "[[File:" string) which is distorted. BTW, I have had a similar thumb problems before - if a PNG is missing the IEND chunk then it is fine full-scale on WP, but disappears in thumb view. Materialscientist (talk) 05:22, 31 December 2009 (UTC)

Well, if the uploaded image is an invalid PNG image, then it's not surprising if Wikipedia can't handle it. But in this case there was nothing wrong with the uploaded image.

I found the methods to purge the image thumbnails: We have to purge the image page on Commons for images uploaded there, it doesn't help to purge the description page here at Wikipedia. And in some cases we need to do an extra trick to the thumbnail URL itself to really purge the thumbnail. Wikipedia:Village pump (technical)/FAQ described one part of the method, and Wikipedia:Purge described another part of it. I have now updated Wikipedia:Purge#For images so it describes the whole method.

--David Göthberg (talk) 06:00, 31 December 2009 (UTC)

Not what links here, but how the link looks

The other day I repaired a couple of vandalism edits where someone had found words like "incompetent" in an article (it was a plot synopsis) and linked them to a well-known public figure. This was a subtle enough form of vandalism, and in an obscure enough article, that it had stood for two months without being noticed.

It occurs to me vandalism patrollers might find it easier to detect these things if they had a tool to do it. To avoid insulting anyone real, I'll make a mild example using "John Doe": Let's say someone took one instance of the word "terrorist" in the article Die Hard and made it a link like this: terrorist. Let's also suppose that John Doe was a high-profile article (but not about an actual terrorist) that people wanted to watch for vandalism to.

What could they do? They could check the "What links here" feature on the John Doe page -- but that would just say Die Hard and wouldn't give any understanding of why it had been linked. Or, as far as I know, of when.

On the other hand, if there was a tool reporting "Details of links to this page" or something like that, it might report

• Die Hard as "extremist"

perhaps with the date as well. And that would be the hint to take action.

Vandals will also make such links the other way around: not terrorist linking to "John Doe", but John Doe linking to "terrorist". If John Doe is a high-profile page, it might be desirable to have a tool that would track links of this kind too. Both tools would be facilitated if there was a database of all links, or all new links, showing the context, the link text, and the target page.

This is just a suggestion in case anyone wants to run with it. If it's already been done, fine. If nobody's interested, fine. If it's technically infeasible, fine. There is no need to reply to me here or anywhere else. --208.76.104.144 (talk) 12:08, 29 December 2009 (UTC)

People occasionally scan the database dumps for vandalism like this. --MZMcBride (talk) 07:02, 31 December 2009 (UTC)

Another media wiki site

I am an admin on another media wiki site. We are getting vandalised by users from 221.206.60.??? Is there a way to block a whole group of IP addresses or does it have to be done one at a time? There is no obvious multiple address option, but I recall folk talking about that feature here. -- SGBailey (talk) 09:46, 31 December 2009 (UTC)

Never had to rangeblock myself, but I gather you would do it by blocking 221.206.60.0/24. Maybe someone else could confirm? I'd hate to be wrong :) - Jarry1250 ^{[Humorous? Discuss.]} 09:57, 31 December 2009 (UTC)

See mw:Help:Range blocks. --Zvn (talk) 10:20, 31 December 2009 (UTC)

Thanks. Another related question - How should a user who has been blocked contact an administrator to explain why they have been incorrectly blocked. If editing has been turned off, then they can't put a message on an administrator's talk page... -- SGBailey (talk) 00:18, 1 January 2010 (UTC)

It depends on how they're blocked. If they can still edit their own talk page, they can use that. Alternatively, if they can use emails, you could have them emailing an admin. Or you could set up a mailing list/email address for block reviews. It's up to you. On enwiki, we use the {{unblock}} template of course. :) Ale_Jrb^talk 00:22, 1 January 2010 (UTC)

Wade_Robson#Routines_for_So_You_Think_You_Can_Dance

Why is that table inside the Personal Life section? Woogee (talk) 21:26, 31 December 2009 (UTC)

Because it wasn't closed. I fixed it for you. Ale_Jrb^talk 21:44, 31 December 2009 (UTC)

what is rectifier?

working and about it —Preceding unsigned comment added by Sripal singh patel (talk • contribs) 08:39, 1 January 2010 (UTC)

Look at rectifier. Rich Farmbrough, 11:05, 1 January 2010 (UTC).

are we can add friend's?

plz tell me —Preceding unsigned comment added by Sripal singh patel (talk • contribs) 08:46, 1 January 2010 (UTC)

Wikipedia is an encyclopedia, not a social networking site. Thus while friendships can and do develop among various contributors to this site there is no technical means to "add a friend" available at Wikipedia. --Allen3 ^talk 11:14, 1 January 2010 (UTC)

Indeed, I've got 41 friends from Wikipedia over on Facebook. But Wikipedia itself is not a networking site, and editors that try to treat it as such can find the site very, very unfriendly very, very quickly. EVula // talk // ☯ // 00:04, 2 January 2010 (UTC)

In other words, no. Equazcion ^(talk) 11:22, 1 Jan 2010 (UTC)

Move unsuccessful

I tried to move WGBT to WPTI but was told I couldn't. I couldn't remember the procedure and the message I got obviously didn't send me to the right place. Then I got distracted and forgot that I needed to find the right place.Vchimpanzee · talk · contributions · 22:42, 1 January 2010 (UTC)

Wait, it wasn't the message. It was the text of the description of how to move. But the correct link wasn't anywhere on the page.Vchimpanzee · talk · contributions · 22:44, 1 January 2010 (UTC)

I'm not sure what you're asking - the history shows that you did successfully complete the move. What's the problem? Ale_Jrb^talk 22:48, 1 January 2010 (UTC)

The message is:

The page could not be moved: a page of that name already exists, or the name you have chosen is not valid. Please choose another name, or use Requested moves to ask an administrator to help you with the move. Do not manually move the article by copying and pasting it; the page history must be moved along with the article text.

Nothing in there told me how to get to the page I finally found.Vchimpanzee · talk · contributions · 22:49, 1 January 2010 (UTC)

Which page are you talking about? If you complete the move, then you don't get that message... Sorry if I'm being silly here. Ale_Jrb^talk 22:57, 1 January 2010 (UTC)

I felt helpless when I tried the links on the page when the move didn't work. I finally remembered there was another page about moves.Vchimpanzee · talk · contributions · 23:05, 1 January 2010 (UTC)

When you first tried to make the move, you correctly got the quoted message because there already was a page at WPTI. You correctly placed {{db-move|WGBT|changed call letters to WPTI}} on the existing page at WPTI (this can only be seen by administrators because it's now deleted). An administrator deleted WPTI and then you successfully made the move. All is well. PrimeHunter (talk) 23:14, 1 January 2010 (UTC)

I'm having a new problem. I click on "What links here" and get a long list of radio station with a template I have already fixed.Vchimpanzee · talk · contributions · 23:09, 1 January 2010 (UTC)

I'm sorry about that. Which page, that finally helped you, do you feel should be added? The 'what links here' page can sometimes take a while to update, especially during peak times, if the links were from a template. Don't worry about them - if you've removed the links, they'll disappear from the list in a little while. Ale_Jrb^talk 23:12, 1 January 2010 (UTC)

I thought it was this one, but I can't find the specifics of the procedure. And none of the speedy deletion pages had it either. I know I saw it today, because I used what was on it.Vchimpanzee · talk · contributions · 23:19, 1 January 2010 (UTC)

I see PrimeHunter's message now. Somewhere I was told to use the procedure he outlined above, and I don't remember where that was. That would be helpful in the message I showed above.Vchimpanzee · talk · contributions · 23:23, 1 January 2010 (UTC)

Hey, wait a minute. PrimeHunter's message will show up if someone searches the Village Pumps. That actually provides one solution.Vchimpanzee · talk · contributions · 23:25, 1 January 2010 (UTC)

The quoted message is from MediaWiki:Articleexists and links to Requested moves which mentions {{db-move}} at Wikipedia:Requested moves#Uncontroversial requests. {{db-move}} is also mentioned at WP:CSD#G6 and Help:Moving a page#Before moving a page but not in the later section Help:Moving a page#Moving over an existing page. PrimeHunter (talk) 23:52, 1 January 2010 (UTC)

It was the uncontroversial requests section. I must not have read the whole thing when I requested the move. Thanks. It appears the directions are adequate, and I just wasn't reading carefully.Vchimpanzee · talk · contributions · 16:56, 2 January 2010 (UTC)

Custom CSS/JS not working on iPod

I'm currently away from home, but when occasionally coming to check up on things here I noted that I saw rollback links on my watchlist, something which I specifically disable through CSS and JavaScript. I know that the CSS and JS work—they've worked in the past, and continue to work if I preview my user JavaScript, but for whatever reason they don't appear to work outside of those test conditions. Does anybody have an idea of what the problem might be? {{Nihiltres|talk|edits}} 23:58, 1 January 2010 (UTC)

You probably enabled other scripts that do not work on the iphone. If scripts fail, they often stop execution of scripts that follow it. Best way to check for an issue here is to use Safari or Chrome, since those browsers use the same renderer (WebKit) —TheDJ (talk • contribs) 14:59, 2 January 2010 (UTC)

That seems unlikely, since my custom CSS is not working either. I'll test against Gadget use when I get back tomorrow, and compare again with the desktop version of Safari, but I haven't noticed any such problems there. I have a feeling that the problem is not one of a failed script stopping other things. {{Nihiltres|talk|edits}} 02:01, 3 January 2010 (UTC)

Subpage count

Is there a way to get a count of subpages, specifically Special:PrefixIndex/Wikipedia:Articles for deletion/? I found [8] on the Toolserver, which outputs a single list – in this case very long, I stopped it around Bi. Flatscan (talk) 07:15, 2 January 2010 (UTC)

I just ran a query on the Toolserver. Pages that match "Wikipedia:Articles_for_deletion/" (not including the main page, obviously):

mysql> SELECT COUNT(*) FROM page WHERE page_namespace = 4 AND page_title LIKE 'Articles_for_deletion/%';
 ---------- 
| COUNT(*) |
 ---------- 
|   206597 | 
 ---------- 
1 row in set (0.23 sec)

This includes non-redirects and redirects (and excludes any pages in the Wikipedia talk namespace). I don't know of a way you could easily generate this number yourself. --MZMcBride (talk) 07:20, 2 January 2010 (UTC)

Thanks much! Flatscan (talk) 07:22, 2 January 2010 (UTC)

What links here

The time that the link was linked should be shown. Post to bugzilla as i have no account.174.3.123.13 (talk) 08:01, 2 January 2010 (UTC)

The index for whatlinkshere, is probably updated on each parsing of a page that includes said link. With page blanking vandalism etc, it would be very difficult to keep accurate logs on when a link was first added, vs. removed vs. pageblanked. Simpler is to log when it was last parsed as being present in the article. But that seems not so useful information to me. —TheDJ (talk • contribs) 14:56, 2 January 2010 (UTC)

Perform AWB's general fixes from a dedicated standalone bot

This is a proposal to make AWB's "general fixes" into a standalone bot whose tasks would be subject to Bot Approval, as opposed to a set of tasks that are not meant to be run as automated, yet certain bots implement anyway. I'm going to try to explain the problem, but it's complicated and my bot experience is limited, so please bear with me.

• Currently, we have AWB, which is software to manually perform assisted edits. It assists editors in editing articles: the editor still needs to sit there and push buttons. AWB has a "general fixes" feature, where it finds and suggests common fixes to a page, shows the proposed edits to the user, and the user can choose whether or not to implement them. In recent times, these general fixes have been deemed so benign and uncontroversial that certain bots started to implement them automatically across all articles (namely User:SmackBot).

• While certain bots may have gotten approval to run AWB's "general fixes", those general fixes change regularly whenever AWB's authors release a new version, and the changes aren't examined or approved by BAG. Ergo unapproved tasks are regularly performed automatically, basically across all articles. These unapproved tasks have been the subject of several recent complaints regarding SmackBot (and perhaps other bots). The author has been deferring responsibility for the edits to AWB's authors, and them back towards him. It seems clear that something needs to be done to fix this.

I see three options:

1. AWB's "general fixes" get written into a dedicated, standalone bot, such as User:AWBGenFixesBot (as described in the title of this section). It will be subject to bot approval. No other bots may perform AWB's general fixes anymore.
2. AWB's "general fixes" simply become subject to WP:BAG approval. AWB's authors would need to seek approval before adding new tasks to "general fixes".
3. A separate online module is created to store AWB's "general fixes" code that has been approved for automated use. It will be subject to BAG approval. This would be separate from, and not affect, the original AWB software. Any bot may use the module, and must use that version of the code if they want to claim to be implementing "AWB's [now-approved] general fixes".

I'm not all that experienced with bots here so I hope this all made sense. Please voice your support or opposition. Thanks. Equazcion ^(talk) 17:46, 31 Dec 2009 (UTC)

I agree with the general direction of what Equazcion said. --Jc3s5h (talk) 18:35, 31 December 2009 (UTC)

Equazcion, I have a natural suspicion of most automated editing (although not of actual bot editing) because editors often don't take responsibility for their automated edits; however, I must ask: this is the solution to what problem? ⇦REDVERS⇨ 18:42, 31 December 2009 (UTC)

WP:BAG exists for a reason. Automated tasks need approval. Here's one example of an improper automated task: imposing a preference for tag placement. Things like this have been complained about and discussed at User talk:Rich Farmbrough in the past. Tasks get added to AWB's general fixes that were never approved and are in fact improper for blanket automated implementation. Equazcion ^(talk) 18:47, 31 Dec 2009 (UTC)

It seems like the problem is SmackBot doing more than the scope of its approval; I don't see the need for a wider solution. With reference to the {{no footnotes}} tag placement, I now see the documentation for that template was updated in August to remove the instruction over its placement. If that had been reported on the AWB bug/feature request/talk page the AWB code could have been updated. I will do it now. Rjwilmsi 19:25, 31 December 2009 (UTC)

The third proposal is far and away the best. It is the solution to two problems, firstly a bot "knowing" that some clean-up is needed and not doing it, and secondly another bot having to come by and do it later (or it remaining undone through being missed or being too minor to be "worth" an edit). Specific example: SmackBot's main task is dating maintenance tags, however any bot (or person) running AWB with gen fixes will successfully date most tags, saving SmackBot from having to do it and saving an edit. Probably this lightens the load a fair bit, and means I can look at the difficult cases much earlier than I would otherwise.

How might this work? Well I'm not an AWB coder but I think something like this:

1. A list of gen fixes kept at WP:AWB/General fix authorisation - a protected page - with auth status.
2. When AWB goes into bot mode it pulls the list and reads it into an array
3. Each GF then has something like "if ((!bot || gfBotAuth.removeSpuriousPipes) && otherConditions)" controlling it
4. BAG is then asked to approve each GF - starting with the non-contentious stuff in a bundle, and the other stuff as consensus can be gained, or as it is added.
5. Any admin can update the page from the BAG approvals
6. Peace happiness and a perfect 'pedia result.

Rich Farmbrough, 20:11, 31 December 2009 (UTC).

A little while ago I suggested this, only I suggested it be done server-side. If that were to be the case, then we could stop bickering about general fixes. Im not sure it can be done. Tim1357 (talk) 16:29, 1 January 2010 (UTC)

I'd certainly be in favour of this, if it was approved in a language-neutral way; I'd like the equivalent for pywikipedia that could be applied by any pywikipedia bot prior to/as part of page save. Josh Parris 23:47, 4 January 2010 (UTC)

Unpatrolled pages

Where did all the unpatrolled pages go? This used to have hundreds of entries, usually going back over a month, but now has just a handful. Are new pages being auto-patrolled in some way? AndrewWTaylor (talk) 11:41, 2 January 2010 (UTC)

Hehe - no. User:Ironholds did a lot (colossally massively awesome amount) of work (with other members of WP:NPP) and cleared the backlog for the new year. Massive thanks to everyone! Ale_Jrb^2010! 11:50, 2 January 2010 (UTC)

It would never go back over a month, because pages older than one month are not shown.. OrangeDog (τ • ε) 11:21, 3 January 2010 (UTC)

Thanks both. AndrewWTaylor (talk) 12:23, 3 January 2010 (UTC)

"Spam protection filter" and Tumblr

I'm being prevented from saving an edit to the Spook Country article due to the spam protection filter not liking an external link to node.tumblr.com, which is essential to the article. The page was saving fine with the link a few minutes ago. Can someone tell me what has happened and what can be done to whitelist the link again? Mahalo,  Skomorokh  12:26, 3 January 2010 (UTC)

Yep, any tumblr urls are blacklisted by the global spam blacklist. You can request a url be removed from the list by asking here. When you say it was fine saving it a while ago, I assume that was without the link, because it hasn't been recently added (as far as I know, no one is exempt from it)... Ale_Jrb^2010! 12:35, 3 January 2010 (UTC)

Actually, I think it would be better to check (or add it to) our local whitelist, unless you think it would be useful to everyone. That's found here. Ale_Jrb^2010! 12:38, 3 January 2010 (UTC)

Thanks for the reply, Ale_jrb; I've filed a whitelist request. I could save edits to the Spook Country article containing the link at 12:18 UTC, but not since 12:21 UTC. I just saved an edit to another article containing the link at 12:43 UTC. However, the filter won't let me add the link to this page.  Skomorokh  12:49, 3 January 2010 (UTC)

In this misformatted edit you removed the link (not the url in the source but the working link in the rendered page). In the following edit you closed a tag which meant the link would have been added where it was not already. The spamfilter prevented that. In the past no page with a blacklisted link could be saved but currently (see rev:34769) the spamfilter only stops addition of blacklisted links. bugzilla:16325 is a request to change the current behaviour. PrimeHunter (talk) 13:00, 3 January 2010 (UTC)

That's very helpful, and solves the mystery; thanks very much, PrimeHunter. A whitelist entry has been added so this particular incident should be resolved happily. I would say that the whole spam filter business is difficult for a regular editor to navigate. Mahalo,  Skomorokh  13:10, 3 January 2010 (UTC)

The blacklist message is from MediaWiki:Spamprotectiontext. You can suggest changes on the talk page. By the way, tumblr.com was blacklisted 10 November 2009 [9] while [10][11] confirms it was already in both the mentioned articles. PrimeHunter (talk) 13:23, 3 January 2010 (UTC)

I may make some suggestions to the message; thanks again, PrimeHunter.  Skomorokh  13:31, 3 January 2010 (UTC)

MediaWiki:Spamprotectiontext is not actually the message I get trying to save tumblr links; it's an angrier version of what I was going to suggest though. The message I get is:

"The page you wanted to save was blocked by the spam filter. This is probably caused by a link to a blacklisted external site.

The following text is what triggered our spam filter: [external links]

Return to [[page I tried to edit]]"

I tried searching through WP and Google but could not find this MediaWiki page.  Skomorokh  13:39, 3 January 2010 (UTC)

As the edit was prevented by the goobal rather than local blacklist, I assume you saw this message. From our local message, though, it sounds like we were expecting that one to be shown for global blocks as well. Hmm... Ale_Jrb^2010! 13:42, 3 January 2010 (UTC)

I just tried to spam my sandbox with a tumblr link, and I got the (correct) local message. I wonder why you didn't. Strange. Ale_Jrb^2010! 13:48, 3 January 2010 (UTC)

Hmmm, I tried it logged out and got the local message. Curious.  Skomorokh  13:54, 3 January 2010 (UTC)

Do you use the en-GB interface language in your preferences by any chance? Or any language that isn't just 'en'? If so, that's the reason I think - I can get the global message by setting to en-GB. Ale_Jrb^2010! 13:59, 3 January 2010 (UTC)

Unfortunatly Links (such as node.tumblr.com) to blogs, personal web pages and fansites are Links normally to be avoided. node.tumblr.com is all three.--Hu12 (talk) 07:36, 4 January 2010 (UTC)

If you have an issue with the link's usage, you should take it up on the article talk page - a respected editor who is working on the article says that its inclusion is essential after all. :P VPT is not the place for the discussion either way. :) Ale_Jrb^2010! 11:41, 4 January 2010 (UTC)

Phantom black square in SVG File:UKExpenditure.svg

Up above in #Why are some SVGs squashy? I got some good advice about fixing the thumbnailing of my SVG by making it all in a font that Mediawiki can render without flipping out. (The text was in Helvetica, I've changed it all to Arial). This has indeed fixed the thumbnailing problem, so the text no longer goes crazy. However, as you can see in File:UKExpenditure.svg, I now have a phantom huge black square in the thumbnail and on the File page. It emphatically does not really exist, as there is no such block in my Inkscape view (not even, say, a transparent block in that place, it just has no existence at all), and it also does not exist if you click through the actual SVG page from the File page. So my questions are:

1. Any suggestions for disposing of The Black Phantom?
2. More abstrusely, I have other files that mix Arial and Helvetica, and they work fine, e.g. File:UK taxes.svg; see the renderings in Taxation in the United Kingdom where its thumbnail is fine, so does behaviour depend on some subtle ordering of the fonts?
3. [Minor rant] How on earth does Wikipedia not support Helvetica, one of the most used fonts in all the world?!

They are SVGs, so feel free to pick over the source. Cheers, Splash - tk 02:13, 3 January 2010 (UTC)

poking through this in XML view on inscape, it turns out that there actually is an object there: you'll see it under g125/g3639/flowRoot3314/flowRegion3316/rect3318 (sorry, maybe not the best way to display the hierachy, but you should be able to find it). --Ludwigs2 02:32, 3 January 2010 (UTC)

Thank you, that fixed it. Though I've no clue how it comes to be there whilst being invisible in all views save Mediawiki! How do I find problems such as that for myself - how did you know what to look for (and do you know why it was ignored until the font was more compliant)? Splash - tk 13:31, 3 January 2010 (UTC)

Open the thing in notepad. Random black squares are a known issues with inkscape although far from common.©Geni 17:50, 4 January 2010 (UTC)

inkscape has a built-in XML editor, and if you select xml blocks in the editor it will outline the object in the main viewer. I just used the brute force method (poking away at likely XML blocks until the transparent block was outlined). --Ludwigs2 19:08, 4 January 2010 (UTC)

as to why empty blocks only show in wikimedia renderings - I have no idea. --Ludwigs2 19:09, 4 January 2010 (UTC)

Wikipedia uses the librsvg render. Inkscape uses a different system.©Geni 19:12, 4 January 2010 (UTC)

Ok, thank you both for the answers. Hopefully one day MATLAB will support SVG export natively and I won't have to be hacking around with the figures anymore! Splash - tk 16:51, 5 January 2010 (UTC)

embedded table. height and width

I need a table embedded in the cell of another table to expand to completely fill that cell. width=100% works fine for width but what do i do about height? Lemmiwinks2 (talk) 20:54, 4 January 2010 (UTC)

text text text text text

top aligned text bottom aligned text

easiest way to do this, I think, is to use style="margin: 0px" on the embedded table, which should set its margins to 0px and force the table to fill the entire cell. --Ludwigs2 21:01, 4 January 2010 (UTC)

I dont think you understand. I already have no margins. For layout purposes i need the inner table to expand to the full size of the cell. Lemmiwinks2 (talk) 21:15, 4 January 2010 (UTC)

Am I missing something, or is there a reason you aren't just using height: 100%? Like so:

text text text text text

top aligned text bottom aligned text

Ale_Jrb^talk 21:49, 4 January 2010 (UTC)

I tried that and it didnt work. I must have screwed it up somehow. Sorry to have bothered you. thanks. Lemmiwinks2 (talk) 22:28, 4 January 2010 (UTC)

I see what i was doing wrong. tricky. I would never have figured it out myself. its working fine now. Thanks. Lemmiwinks2 (talk) 22:36, 4 January 2010 (UTC)

No problem. Glad you got it fixed. :) Ale_Jrb^talk 16:06, 5 January 2010 (UTC)

More on my history freezing problem

I've discovered that when I'm logged out, I don't have the problem with history pages freezing. So there must be something in my setup that's causing this problem, but I have no gadgets turned on. Any ideas? Woogee (talk) 21:55, 4 January 2010 (UTC)

The earlier section about this is at Wikipedia:Village pump (technical)/Archive 69#Is nobody else having problems with viewing history pages? You could try disabling any browser addon or toolbar you might have, or changing other things than gadgets at Special:Preferences. PrimeHunter (talk) 22:33, 4 January 2010 (UTC)

I reset all of my preferences to the defaults, and the problem went away. I don't know what caused it. Woogee (talk) 01:51, 5 January 2010 (UTC)

Possibly, you were one of the users who was being followed with clickTracking by the usability team... I remember that those people were selected randomly, and I think that they were tracked for a while (cookie ?) that might explain why a history page would become slow (awful lot of form elements to track I guess). —TheDJ (talk • contribs) 11:14, 5 January 2010 (UTC)

questionnaire

radio buttons please

Is it possible to create an on-wikipedia questionnaire with radio buttons. A questionnaire in which information can be saved and compiled? Ikip 03:42, 5 January 2010 (UTC)

Not on Wikipedia, but here is a MediaWiki extension that probably does what you want. Graham87 04:35, 5 January 2010 (UTC)

Yeah, a more sophisticated version is mw:Extension:Quiz, neither save data to my knowledge, and neither are loaded on wikipedia. Ikip 08:51, 5 January 2010 (UTC)

I know this is archived, but wikia uses a Poll: http://help.wikia.com/wiki/Polls

Full extension information is found here: http://dev.wikia.com/wiki/Extension:AjaxPoll Adamtheclown (talk) 18:25, 17 November 2010 (UTC)

Wikipedia:Mobile access

Whilst trying to find info on using Wp on mobile devices I find myself in the realms of lack of info, Wikipedia:Mobile access briefly mentions en.m.wikipedia.org, promptly proceeds to slate it and then lists many other third party portals - I can find no other mention or link to anything about this. Is en.m.wikipedia.org an official part of wikipedia - and does it have a central point of reference discussion? The current article looks like an anti-advert for it! Lee∴V (talk • contribs) 13:04, 5 January 2010 (UTC)

en.m.wikipedia.org is an official wikimedia portal yes. And that page is indeed heavily outdated. —TheDJ (talk • contribs) 13:42, 5 January 2010 (UTC)

Orange links revisited

I asked the question here. One answer I didn't get is why, when I test links while previewing what I am about to add to Wikipedia, the NEXT link to test is the one that is orange, not the one I already tested.Vchimpanzee · talk · contributions · 22:58, 5 January 2010 (UTC)

An active link (which your browser colours orange) is the color of a link when it has focus, either by clicking on it or tabbing to it. It has nothing to do with whether you have visited a page or not. OrangeDog (τ • ε) 13:25, 6 January 2010 (UTC)

Search box help

I added a search box to one of my user pages, but searches seem to return no results. Does it take time for the pages to be crawled or added to the fulltext index, or should searches work immediately? SharkD  Talk  06:53, 6 January 2010 (UTC)

Yes, it takes a while for recent changes to be reflected ion searches, see Wikipedia:Searching#Delay in updating the search index. The search of your subpages seems to be working for me now ([12]). Svick (talk) 08:42, 6 January 2010 (UTC)

Awseome, thanks! SharkD  Talk  05:44, 7 January 2010 (UTC)

rowspan and columnspan and table sorting?

Evidently it's not possible to have a table automatically sort if there are any rowspan or columnspan values in it Help:Sorting. Isn't there some way of enabling the software to make that possible? Is the only workaround to get rid of the spans and duplicate words? Шизомби (Sz) (talk) 00:22, 7 January 2010 (UTC)

I think others have tried... bugzilla:8028 http://tablesorter.com might be a potential solution if we go fully jQuery in the future (but that will be a while). —TheDJ (talk • contribs) 12:21, 7 January 2010 (UTC)

The styling of those tables is a lot nicer than ours, too. SharkD  Talk  19:51, 7 January 2010 (UTC)

Disabled animated gif (since yesterday)? (software change?)

Resolved

 – Not fixed yet, but bugzilla:22041 covers it. Thanks, Dragons flight -- Proofreader77 ^(interact) 19:48, 7 January 2010 (UTC)

Has something happened code-wise in the past day.

(1) An animated gif has frozen. (2) And some "div" absolute placement is now messed up. [that problem went away]

Re: #1

NOTE: http://en.wikipedia.org/wiki/File:Phenakistoscope_3g07690a.gif should be an animated gif

And it is so if you click down in the upload table. http://upload.wikimedia.org/wikipedia/commons/a/a2/Phenakistoscope_3g07690a.gif

The freezing has taken place in the past day.

re (2): I'd carefully aligned some things up top, but now they've moved higher. See: http://en.wikipedia.org/wiki/User:Proofreader77 (There was not overwriting yesterday.)

Note; I see Foxfire has updated, abut I tried with Google Chrome also, and the problems exist there too.

-- Proofreader77 ^(interact) 00:45, 7 January 2010 (UTC)

As currently configured the software is supposed to show only the first frame of an animated GIF if the original height and width are such that height*width > 1 million or height*width*frames > 12.5 million. The limits are there to avoid resource overruns. However, it appears that it is freezing a variety of images even when these limits aren't reached so there is probably a problem in the limiter calculation. (Actually, the first of these can probably be removed now given the recent improvements in the GIF handler.) Dragons flight (talk) 10:28, 7 January 2010 (UTC)

Do I (or someone) need to make a smaller animated gif of http://en.wikipedia.org/wiki/File:Phenakistoscope_3g07690a.gif

Reminder - the copy at "upload." ... animates fine. http://upload.wikimedia.org/wikipedia/commons/a/a2/Phenakistoscope_3g07690a.gif

(Note: the still frame display only began yesterday. It had been animating fine until then.) Proofreader77 ^(interact) 12:19, 7 January 2010 (UTC)

There was a configuration change 3 days ago. I don't think you ought to need to change it because I believe it is a server side mistake since the image does not appear to breach the limits. Dragons flight (talk) 12:35, 7 January 2010 (UTC)

bugzilla:22041 Dragons flight (talk) 13:05, 7 January 2010 (UTC)

Many thanks. Looks like issue is well documented for fixing. So I can just sit back and relax. :-) Proofreader77 ^(interact) 19:48, 7 January 2010 (UTC)

CentralNotice glitch on IE8

When logged out, I am receiving an approx. 12-px high, empty light gray box in the place where CentralNotices usually appear. The page code suggests it is to do with a "global sysops" notice. This is under IE8; it does not appear in Firefox 3.5 - it must be one of IE's many CSS glitches. — This, that, and the other (talk) 06:36, 7 January 2010 (UTC)

I can't find it in IE8 at the moment. Maybe it's been fixed? Probably just a hangover from the fundraiser css hacks. - Jarry1250 ^{[Humorous? Discuss.]} 10:24, 7 January 2010 (UTC)

Oh, I see, it relates to meta:Global sysops (for which a central notice is now appearing on Commons). Will probably be with us soon too. - Jarry1250 ^{[Humorous? Discuss.]} 10:54, 7 January 2010 (UTC)

Special:Book

The Book namespace has been created, but this maintenance template page Special:Book still saves books on the old page structure, Wikipedia:Books/ (on the bottom right of the screen, in the "Save and share your book" box). Any idea where I can best request a change to this? Fram (talk) 10:15, 7 January 2010 (UTC)

I asked in the pediapress IRC room, apparently that is one of the few strings that isn't configurable in the collection extension :D It will almost certainly require a codechange, and they are looking into it now. —TheDJ (talk • contribs) 13:20, 7 January 2010 (UTC)

Thanks! Fram (talk) 13:31, 7 January 2010 (UTC)

Ticket: http://code.pediapress.com/wiki/ticket/760 —TheDJ (talk • contribs) 18:06, 7 January 2010 (UTC)

Page notice links

We have now added new features to the editnotice system: As most of you soon will notice, when you edit pages you will now sometimes see red or blue links at the top of the edit page looking like this:

Page notice

Or like this:

Group notice   Page notice

Those links are to the group and page editnotices of the page. Admins and accountcreators always see both links, even if the notices have not yet been created, since they can create and edit the editnotices. Normal users see the red "Page notice" links only on user and user talk basepages, since they can create and edit such notices. Normal users also see blue links to group and page notices if they already have been created, so they can find and view the source code of the notice. But they still can't edit those notices, except the ones in userspace.

To learn more about editnotices and to discuss these new features, see Wikipedia:Editnotice and its talkpage. Oh, and we would love to get some feedback about what people think about the editnotice system.

--David Göthberg (talk) 22:19, 7 January 2010 (UTC)

Infobox at Grand Blanc Community High School

What's wrong with the infobox at Grand Blanc Community High School? Woogee (talk) 03:25, 8 January 2010 (UTC)

I fixed up the athletics infobox. Still needs a school infobox. See other school articles for examples.LeadSongDog come howl 04:39, 8 January 2010 (UTC)

{{mtc}}

Resolved

 –  – ukexpat (talk) 16:47, 8 January 2010 (UTC)

Will somebody please help me merging these 3 images to Wikimedia Commons? I can not do it myself.

[[File:Rotunda interior steinway hall nyc mia laberge art case piano.jpg]]
[[File:Artist mia laberge at kennedy center unveil of art case steinway.jpg]]
[[File:Artist mia laberge with henzy z steinway.jpg]]

At Commons there is a category named Steinway & Sons. The 3 images can be added to this category.

Thank you.

done. bit of a niggly procedure; I should look into that. but your files should be on commons under the same names. let me know if there's any issues.

I have to say, grand pianos are truly beautiful objects. form and function, tied in a knot. --Ludwigs2 11:17, 8 January 2010 (UTC)

Thank you so much. It is all very good. Fanoftheworld (talk) 11:56, 8 January 2010 (UTC)

Convert template help needed

Resolved

 – finding the right redirect fixed the problem Sswonk (talk)

Is anyone who watches this page familiar enough with {{convert}} to help me with an apparently simple but esoteric solution to an issue? Specifically, I want to convert acres to hectares but display the output number only for use in a table with the units given in the header, i.e. "241 (98)" instead of "241 acres (98 ha)". To do so, I used the syntax 241 ({{Convert|241|acre|ha|disp=output number only}}). That resulted in a redlinked subtemplate being displayed. I created the redlink as a redirect, see Template:Convert/LoffAoffDoutput number onlySoffNa, just to see what would happen. This is not easy to do with a sandbox as the main template can't be easily edited for testing. I need to fix this, because as of now it just outputs the conversion factor, i.e. "241 (98)", no matter what the input. Sswonk (talk) 18:33, 8 January 2010 (UTC)

I redirected Template:Convert/LoffAoffDoutput number onlySoffNa to Template:Convert/LoffAoffDoutput number onlySoff instead of Template:Convert/LoffAoffDoutput number onlySoffT like you did and it seems to work now. Svick (talk) 19:43, 8 January 2010 (UTC)

Thank you, that did it. The system of convert subtemplates is vast, so getting as far as I did gives me some measure of hope for future endeavors in that realm. This one's done, thanks again. Sswonk (talk) 19:51, 8 January 2010 (UTC)

Galleries

Help:Table#Wikitable_as_image_gallery explains advantages of using a table over <gallery>.

Wikipedia:Extended_image_syntax#Compatibility_considerations and Wikipedia:When_to_use_tables#When_tables_should_not_be_used says that tables should not be used for these purposes.

This is probably not possible with <gallery>. I suggest making this a priority for bugzilla to implement.174.3.101.61 (talk) 04:26, 9 January 2010 (UTC)

SVG Skin

Is it technically possible to have a MediaWiki skin which not only uses SVG for the interface but also passes SVG files directly to the browser instead of rendering them as PNG first? This would be really cool for large monitors, the visually impaired and anyone who surfs this site while zoomed in. I understand there would be some issues with unoptimized images hanging the browser, but if Wikipedia made the push to a vector-based web, we'd likely see WebKit and Mozilla deal with these problems in browserland Bitplane (talk) 14:37, 8 January 2010 (UTC)

There are several problems

• For content, SVG files can be much larger than small pngs
• It is VERY difficult to detect how well a browser supports SVG. Especially for anonymous users, we would have split the caching of pages into "people who can handle SVG" vs. "people who cannot handle SVG" and all the detection for that has to be done as well. That's simply not possible
• So the only thing I can see as a possibility is an SVG skin, that people can select themselves in the preferences. But i don't think that does much good for support.

See also (outdated page) m:SVG image support —TheDJ (talk • contribs) 16:14, 8 January 2010 (UTC)

I agree that they can be larger, but well formed ones usually aren't. I agree that it should also be a user preference rather than automatically detecting Chrome, Firefox and Safari users and just enabling it (though this would work in the majority of cases). If I did the work and created such a skin, is it likely to be accepted? Bitplane (talk) 19:19, 9 January 2010 (UTC)

No idea, I don't speak for the developers. You can submit patches to bugzilla: though. —TheDJ (talk • contribs) 19:25, 9 January 2010 (UTC)

I do not see any technical problems with creating a new and separate SVG skin-fork, preferably from vector. But I do not see the big improvement that would justify such a major work with creating and maintaining that skin. The Wikipedia user interface is mostly text based and does not rely much on images or icons. Cacycle (talk)

Special:UnwatchedPages

Special:UnwatchedPages is giving me absolutely no results. Am I to take it that every article is now watched? I'm an admin, so whatever is happening is not an issue with permissions. Nyttend (talk) 00:27, 9 January 2010 (UTC)

As UnwatchedPages is a protected page, normal users are not going to be able to help out. Try posting at AN. 124.168.93.126 (talk) 06:29, 9 January 2010 (UTC)

For the record, its empty for me too. Don't know why, but its not just you--Jac16888^Talk 14:16, 9 January 2010 (UTC)

This has been disabled for a long, long time afaik, so that vandals can't easily find pages that have nobody watching them. You can find the exact number of watchers for a specific page by using this handy tool. It only works for pages with over 30 watchers, however: it won't tell you whether there are 29 people or 0 people watching a page though, for the same reason as above. Ale_Jrb^talk 15:40, 9 January 2010 (UTC)

It was always available to admins, I don't remember hearing about it being disabled for us too--Jac16888^Talk 15:42, 9 January 2010 (UTC)

Meh, I never realised that. Weird then. Ale_Jrb^talk 15:45, 9 January 2010 (UTC)

Bot problem

Moved from Wikipedia talk:Village pump#Bot problem. Svick (talk) 16:46, 9 January 2010 (UTC)

Hi there. There is a bot doing strange things. I can't seem to find the shut-off. Here is an example, and contribs. Sorry if this is the wrong place to post this. [13] [14] Anna Frodesiak (talk) 00:14, 9 January 2010 (UTC)

Actually, it's just the signature that doesn't seem to render properly. Maybe it's just my display. Maybe the bot has already been turned off. Cheers. Anna Frodesiak (talk) 00:17, 9 January 2010 (UTC)

It's not just your display. The (protected) page to turn this bot off is User:SQLBot-Hello/welcome.run and it was already done, so this shouldn't be a problem anymore. Svick (talk) 16:50, 9 January 2010 (UTC)

please refresh category-list articles regularly to reflect deletions

My apologies if this is a worn-out subject. First time I ever thought about it, was today.

I viewed the (apparently, auto-generated) article about people with Asperger's. It included a "Nicky Reilly" link. But, the word Asperger's does not appear in today's version of the "Nicky Reilly" article!

Seemingly, at one time Asperger's was mentioned at the tail end of the Reilly article, in the Categories box. And seemingly, the Asperger's entry was subsequently removed from the Reilly article; and seemingly, there is not a regularly scheduled, automated re-generation of the "people with Asperger's" article, that would reflect the deletion in the Reilly article.

Naturally I suppose this means that there is not an automated, regularly scheduled re-generation of other category-list articles, either. It would be good to have a regular synchronization job like this.

But, all of this is conjecture on my part. —Preceding unsigned comment added by Publius3 (talk • contribs) 21:07, 4 January 2010 (UTC)

Nicky Reilly has no own article. It is a redirect to 22 May 2008 Exeter bombing. The redirect page [15] is in Category:People with Asperger syndrome. It is displayed in italics in the category. This means it is a redirect. PrimeHunter (talk) 22:41, 4 January 2010 (UTC)

Thank you, PrimeHunter, for responding...but, after re-checking, I still do not see Asperger's mentioned there. Yes, the Nicky Reillly article to which I referred is the Exeter bombing article. And yes, there is a Category box in that article. But, I've cut and pasted the present contents of that Category box to here: 'Categories: Terrorist incidents in 2008 | Terrorism in the United Kingdom | Failed terrorist attempts | 2008 in England | Exeter' ...and as you can see, there is no mention of Asperger's. Should I be looking somewhere else? I am puzzled. In support of my speculation that Asperger's may have been first added to, and then removed from, that box, I see that the edit history mentions Asperger's in July 2009. But there is no longer any mention of Asperger's in the Exeter bombing/Nicky Reilly article itself. Publius3 (talk) 11:39, 7 January 2010 (UTC)

Look here. Lupo 12:03, 7 January 2010 (UTC)

Somebody removed the category yesterday so see the old version here instead. If you click on a Nicky Reilly link or enter it in the search box then you are automatically taken to the target of the redirect where the Asperger category never was. It is the redirect page itself which was in the category. My first post linked to it in "The redirect page [16]" (but that link is to the current version without the category). PrimeHunter (talk) 10:32, 10 January 2010 (UTC)

How do I add a Wikipedia template to my own wiki?

Sorry if this is in the wrong section. There's nothing I can see on Mediawiki that tells me how to do this. I'm a fan of Template:Quote_box and would love to use it on my own wiki. I'm not seeing a similar extension on Mediawiki and Cite and Poem don't quite do the same thing. So my question is, how do I use this template on my own wiki? (Assuming I can.) Thanks in advance. Peter Greenwell (talk) 15:09, 9 January 2010 (UTC)

I think you just copy this--Jac16888^Talk 15:24, 9 January 2010 (UTC)

Did that, but it does this --> http://img215.imageshack.us/img215/7377/quotebox.gif As you can see, it formats the quote fine, but it displays the CSS and some of the coding. I'm assuming I need to add more or enable something further for the formatting to come out right and for the coding to be invisible. Thanks for your help so far. Peter Greenwell (talk) 15:41, 9 January 2010 (UTC)

try creating this one too, [17] if you haven't already--Jac16888^Talk 15:44, 9 January 2010 (UTC)

Also [18] and [19]--Jac16888^Talk 15:45, 9 January 2010 (UTC)

It looks to me like you don't have mw:Help:Extension:ParserFunctions installed (or they're not working). Do you? Ale_Jrb^talk 15:48, 9 January 2010 (UTC)

Could be, my guess was just that most templates use other templates in long chains--Jac16888^Talk 15:49, 9 January 2010 (UTC)

That's true, but from the source of the quote box template, none of those are used. And note that the only code that's displaying is that inside the parser function sections. Ale_Jrb^talk 15:51, 9 January 2010 (UTC)

I do have parserfunctions installed (in extensions/) and the proper line in LocalSettings.php there. Funny thing is that Special:Pages Version isn't showing it. In other words, I don't think it's being loaded, which is an entirely different problem I'll need to head over to Mediawiki for. Thanks again for all the help. Peter Greenwell (talk) 16:03, 9 January 2010 (UTC)

Just let you guys know I got it working. My server was reading a cached version of LocalSettings.php, not the newer one I changed. It's all good now. Peter Greenwell (talk) 03:05, 10 January 2010 (UTC)

Good. :) Ale_Jrb^talk 11:07, 10 January 2010 (UTC)

bits.wikimedia.org?

What's up with the new subdomain? Is there anyplace to specifically discus the javascript that we're using (aside from here)?
— V = I * R (talk to Ohms law) 11:47, 10 January 2010 (UTC)

bits. is a new domain that serves content that is mostly static (such as skin JS and CSS, centralnotice). Some ideas behind it are detailed here —TheDJ (talk • contribs) 12:46, 10 January 2010 (UTC)

Thanks. I really with that the decision making process to make these sorts of changes here on Wikipedia were more transparent.
— V = I * R (talk to Ohms law) 13:24, 10 January 2010 (UTC)

It's perfectly transparent. It just is a lot of information, and keeping up to date with everything requires multiple day time jobs (which is why the foundation has 30 people employed these days) —TheDJ (talk • contribs) 13:32, 10 January 2010 (UTC)

Just from looking around, I guess that you're an "insider" in the MediaWiki developer community. As an editor here, and as someone with quite a bit of experience with both private and open source development, I've got to tell you that the developer community is awfully opaque here. Based on your statements and those by others I'm guessing that isn't the intent, but I'm here to tell you that the image presented to the English Wikipedia community (those of us who are not on the inside of the MediaWiki dev community, at least) is that development decisions are made in a "smoke filled back room", and that wider input is generally unwelcome.
— V = I * R (talk to Ohms law) 14:35, 10 January 2010 (UTC)

I'm sure the devs would love more outsider interest, it's understanding enough to feel able to contribute that's the problem I find (personally). Also, Ohms, you have to be careful: to a large extent (as I understand the change) this wasn't a MediaWiki decision, this was a Wikimedia decision - largely a similar set of people, but by no means the same. - Jarry1250 ^{[Humorous? Discuss.]} 14:59, 10 January 2010 (UTC)

Indeed. When Wikimedia decides that performance of the server park is benefited with a dedicated server with specialized caching to handle some static content shared by all installations, then that hardly concerns you as a user (Imagine if facebook were to notify you each time they added/removed a server from their serverpark). The information is out there, you just need to know where to look. And I'm not an insider, i'm just an editor interested in the tech side of wikipedia, who knows where to look when people ask questions like your original question. —TheDJ (talk • contribs) 15:20, 10 January 2010 (UTC)

Most development decisions are only made by the person who commits the change. Then every change is reviewed by a senior developer before it goes live on the site. Though as was pointed out, this was not a MediaWiki decision, nor one that affects the average end-user. But I'm not sure how much more open the development community can be. If anyone is an outsider, its only because they haven't gone inside yet. mw:Communication and mw:Developer hub are both linked to on the main page of MediaWiki.org and list the primary modes of communication for the developers, the IRC channel and the mailing list, both of which are open to anyone. Bugzilla and CodeReview are also open to anyone. Mr.Z-man 15:46, 10 January 2010 (UTC)

I'm confused on where this so called 'back-room' decision complaint is coming from. Nothing changed except the address where things, users have no control over, are being hosted from. It's like getting angry of them changing what web server serves up the pages. Q ^{T C} 15:55, 10 January 2010 (UTC)

Frankly, I don't care about your interest. I'm a volunteer too, and I publish as much information as possible. Unfortunately, you will have to come and get that information yourself, I won't deliver it to you. You are not special. Thanks! Bye! Domas Mituzas (talk) 16:14, 10 January 2010 (UTC)

Bits.wikimedia.org taken out of rotation due to reachability issues. Under investigation. —TheDJ (talk • contribs) 15:57, 10 January 2010 (UTC)

Dear enwiki community,

We would like to remove the server 'srv56' from Apache rotation. That means this server will no longer be serving pages to users. Can someone please describe the approval process we need before doing this? Do we need consensus from the community for such a change or is a majority vote sufficient? Are there any enwiki guidelines we should follow when making this change?

Thanks! kate.

Kate, you are such a troll! We all know that you want to take out two servers, not just one! 16:02, 10 January 2010 (UTC) —Preceding unsigned comment added by Midom (talk • contribs)

Sure, be derisive and dismissive, as usual. *rolls eyes* I'll simply continue to not even try to volunteer my expertise to the development community here. If IRC and mailing lists are the definition of "openness" here, that speaks volumes by itself.
— V = I * R (talk to Ohms law) 16:20, 10 January 2010 (UTC)

OMG, we just lost some engineering talent! Ring the alarm bells! What information didn't you get on IRC, btw? Which channel did you try? #overflow? Domas Mituzas (talk) 16:39, 10 January 2010 (UTC)

This is nothing to do with development community or software. This is a Wikimedia specific change made by system administrators, and I really can't think of a reason users would want to be consulted, or even care. (Or do you think we should not make any technical changes before asking users if it's okay?) kate.

We use mailing lists and IRC for communication, a bug tracker for tracking bugs, and a few wikis (mediawiki.org, wikitech.wikimedia.org) for documentation. These are the standard media for communication among open-source developers. They're generally open to the public and publicly logged (especially the dev stuff, sysadmin stuff is less open). I don't know what you think is more "open" than that – are developers to use the English Wikipedia for communication? That's outside enwiki's scope. Wikimedia runs a thousand or so wikis, and developers aren't going to single out one or two of them to talk to about technical changes. We have our own lower-volume media, and interested parties can subscribe to them.

A remarkable number of enwiki users seem to think that developers and sysadmins exist for the purpose of serving enwiki editors. We don't. Those of us who are paid are paid by Wikimedia, which has a mission that reaches far beyond enwiki and isn't accountable to enwiki. Those of us who are unpaid have our own reasons, which might or might not include helping enwiki editors, and almost certainly does not include helping any particular subgroup of editors. Asking us to talk to you about changes we make that barely affect you anyway, when you haven't even made a minimal effort to involve yourself in what we do, would be like an anonymous Wikipedia reader getting upset because you didn't consult them on a change to WP:NPOV. You want to be involved, you involve yourself. —Simetrical (talk • contribs) 17:51, 10 January 2010 (UTC)

In the interests of drawing this conversation to a close, would it be feasible to make (or is there already) an RSS stream that posts blurbs on these kinds of technical issues? That would solve the supposed transparency issue without generating a whole lot of mess. --Ludwigs2 22:15, 10 January 2010 (UTC)

There's http://wikitech.wikimedia.org/view/Server_admin_log and http://wikitech.wikimedia.org/index.php?title=Special:RecentChanges&hidebots=0 - probably with RSS versions somewhere. Domas Mituzas (talk) 22:58, 10 January 2010 (UTC)

Which is relayed over Twitter/identi.ca/IRC. pick your poison you could say. —TheDJ (talk • contribs) 23:01, 10 January 2010 (UTC)

User page protection

What is the default protection level for User pages? I have a series of User pages of a statistical nature that I feel might be prone to tampering. Thanks. SharkD  Talk  05:43, 7 January 2010 (UTC)

I believe user pages and user subpages are editable by anyone, even anons, except for userspace .js and .css files (like User:Example/monobook.js and User:Example/monobook.css), which are editable only by the "owner" user (and admins, possibly). If you have pages in your userspace that you don't want tampered with, just watch them closely and revert any unwanted changes. — This, that, and the other (talk) 06:52, 7 January 2010 (UTC)

[And admins.] You can also request protection, but this is only done after a proven record of vandalism. [For completeness, I should note that user talk-page protection is very rare for obvious reasons.] - Jarry1250 ^{[Humorous? Discuss.]} 09:59, 7 January 2010 (UTC)

You can have any content you like in a user page, even if it ends in the suffixes ".css" or ".js", so if you have a user page that you want to protect, move it to a title ending with .css or .js. Graham87 12:47, 7 January 2010 (UTC)

The problem is that I'm not active enough to monitor the pages on a daily basis, and the pages are meant to be used as a reference resource for editors, so factual accuracy is critical. Given the amount of noise that occurs in articles like this, I think the protection is warranted. SharkD  Talk  19:45, 7 January 2010 (UTC)

Provide people with a link to a specific revision of the page, or disclaim that people should check the history before they use the information. No-one should be citing an on-wiki page anyway, but independently checking the actual sources themselves. OrangeDog (τ • ε) 21:59, 7 January 2010 (UTC)

I would suggest moving it to the Wikipedia:WikiProject Video games, you even list this as a See Also, perhaps as Wikipedia:WikiProject Video games/Top video games lists. Then it could be monitored/maintained by multiple people, and as discussed above, considered for protection if a true rash of vandalism occurs. Great resource! Cander0000 (talk) 02:55, 10 January 2010 (UTC)

I suggested this on the project's Talk page, but consensus there seemed to be against this. SharkD  Talk  17:19, 11 January 2010 (UTC)

subst templates

Is it possible to make template, which will work either as plain template and as substitution? For example,

• {{subst:review}}

should produce

• {{review|7 January 2010}}

Also, using without parameters:

• {{review}}

should produce different output (error message). A^VB^talk 15:50, 7 January 2010 (UTC)

It would be far easier just to use two templates - one that you subst, and one that it calls. PROD works like this (subst:prod fills out the dated prod template). Ale_Jrb^talk 16:11, 7 January 2010 (UTC)

Yes, you can have different output depending on whether a template is substituted or not, see {{Error:must be substituted}} and {{Error:must not be substituted}} for starting points. However, if you want to require substitution, there is no way to use this trick without leaving artifacts in the expanded template (typically at least an empty #if with an error message).
Amalthea 16:15, 7 January 2010 (UTC)

• use two templates - but this makes more headache for users, which should now remember one more template. artifacts in the expanded template - if you mean "non-substituted template", then this is not too worse if these artifacts will not too noticeable. May you provide bones for such template? {{Demo of template that must be substituted}} is not suitable, because my template is more complex, than plain text string. A^VB^talk 14:37, 8 January 2010 (UTC)
  • No, it doesn't. They only have to remember the subst template - that one handles calling the non-subst. Ale_Jrb^talk 15:32, 8 January 2010 (UTC)
    • it doesn't - ...in theory. PS: Another reason: I not like to produce extra unnecessary pages. PPS: It's pity that current language not support subroutines. A^VB^talk 16:36, 8 January 2010 (UTC)
      • As they still only have to remember one template, it doesn't. End. And a single extra page is much better that having left-over parser code on the page every time the template is used. Ale_Jrb^talk 16:39, 8 January 2010 (UTC)
  • Well, now that I've looked again at what you actually want to do: The best you can do is that the smart, substed template will expand to something like:
    {{#if:|{{review/core||Foo|{{{date}}}}}|{{review||Foo|date=January 2010}}}}
    That will look the same as {{review||Foo|date=January 2010}}, but the page source will contain the additional garbage, and for tags that's pretty much unacceptable. Avoiding that would require a change to MediaWiki. Like Ale_jrb said, that's why numerous similar templates solve this with an additional helper templates, like {{subst:prod}}, {{subst:orfud}}, or {{subst:nfurd}}.
    Amalthea 16:57, 8 January 2010 (UTC)
    • I was thinking, may instead of literally "{{subst:review}}" you could use {{review|subst:}} but I can't think how that would work in your context vis-a-vis junk artifacts - probably more, not less :) - Jarry1250 ^{[Humorous? Discuss.]} 17:11, 8 January 2010 (UTC)
      • That would have to be {{subst:review|subst:}}, but you're right, I forgot that alternative (because I hate it).
        A proper solution (which would require a MediaWiki change, as I said above) would be to have an automatic parameter, like {{{__SUBST__}}}, which automatically expands to "subst:" if the template is being substituted, and is empty otherwise. And while one is at it, a {{{__NAME__}}} which expands to the current templates name (not the page name where it's used) would be very useful for navboxes, where we currently need to tell the template its own name. Amalthea 17:50, 8 January 2010 (UTC)

Yes it would, there are masses of navboxes where this is set to an outdated or wrong value. Rich Farmbrough, 17:02, 11 January 2010 (UTC).

• require a change to MediaWiki - pity. :( A^VB^talk 00:30, 9 January 2010 (UTC)

Such a change has been requested, and turned down, on the reasonable basis that the whole idea of subst is that it produce exactly what the template does in aspic. And I have spent a considerable amount of time trying to do what you are asking for with results that come down to what Amalthea says - and which amount to exploiting a bug, which is bad because either the bug needs to be sustained or the exploit breaks. The only way of cleanly doing what you want is to have (with the correct brackets) #if{1|}||[Category: Reviews that need dating] and use a bot to add the dates - or similarly use a bot to remove the artefacts - you will see code in most of the maint tags to alert SmackBot when these have been substed and need de-substing. Rich Farmbrough, 16:56, 11 January 2010 (UTC).

• use a bot - unsuitable in my case, because template should show link with date in it. A^VB^talk 23:23, 11 January 2010 (UTC)

testing templates

I was wondering if Wikipedia has a "testing option" for testing heavily used templates that have been modified? This would be something that a group of users could access to test new templates that reside on a "test space" on wikipedia. This option would assure that for any template found by the wikipedia software when reading a page, that the test space would first be checked to see if this template resides there. Then if the template was not there, the main template space on Wikipedia would be searched. This would be a step between a user testing a template in their own user space and actually loading it into the live wikipedia -when the user felt the template was ready, they would copy it over to this test space and ask a group of experienced users to check out the template before going live. This would allow a group of users to call up different wikipedia pages using this template and check if the template in the test area affects the many different ways of calling a heavily used template. When you test a template in your user space, you have to dummy up a lot of stuff for testing. With an option like this, a group of users could test the use of a new template without having to dummy up anything. Is there anything like this? stmrlbs|talk 18:04, 9 January 2010 (UTC)

The usual way of dealing with changes to complex templates is to use template test cases. This way, anybody can change the sandbox and then see how would it affect the template output in one place. I think this way is better than your proposal, because even if I'm interested in templates, I want to test one template at a time not all templates that have some untested changes and it's better to have one place for each template, where rare cases can be tested, rather than going through few real pages, finding that template there and ensuring it's okay. Also, in testcases you can easily compare how original and changed versions of a template behave. Svick (talk) 18:26, 9 January 2010 (UTC)

I am aware of how to test templates in a user space. But, like I said, in order to test it, you have to dummy up any pages that call the template, too, in order to test it. There is no way to get a listing of all the different ways a template can be called - you can't search for all the calls of a template where a certain parameter is used a certain way. I understand what you are saying about calling in other people's templates from a test area, but a testing option could also be set up to specify a variable testing area (perhaps a sub user space). The advantage to testing with something like this is that you could test by browsing any wikipedia page that calls the template without having to create a "dummy page" to specifically call the test template. This is a method used in many other systems for just this purpose. I was just wondering if Wikipedia had something similar. stmrlbs|talk 18:36, 9 January 2010 (UTC)

When I'm testing a change to a complex template, I do the following

1. make a copy of the template in my userspace, and make the changes there
2. go back to the original template, click "What links here"
3. open a randomish selection of the pages that transclude it (usually trying to get some from redirects as well) - maybe 5 to 10, depending on how big the change is.
4. on each of those pages, I open the section that uses the template in edit mode and change the template name to my userpage mockup
5. then I use preview mode to test the mockup - I can just leave the page in edit mode, repeatedly clicking preview to test revisions to the userfied template, change and add and remove parameter values to see what happen, and etc. I just have to remember to click close when I'm done so I don't commit any change to the page itself.

this obviates the need to mock up transcluding pages, gives a good selection of real-world uses of the template to test on, and allows me to clean up all the more obvious errors in coding. I don't see any advantage to having a specific 'testing mode' or 'testing option' that preview mode can't handle. --Ludwigs2 19:15, 9 January 2010 (UTC)

this works fine if the template is not heavily used, and you know a good variety of pages to do this "open, edit, preview" method to test, and this is probably a good way to test any template modifications. But for a heavily used template (like infobox road template) that is called by 10,000 pages, and has multilevel calls with different parameter variations with each nesting level, this isn't enough, imo. It would be nice for any template that is heavily used to be able to be able to be tested by a group of users on a project (like the wikipedia road project). I was thinking with a site as big as wikipedia, that perhaps there was some mechanism in place to do this that I didn't know about, but I guess there isn't. Just thought I would check. stmrlbs|talk 19:30, 9 January 2010 (UTC)

I've worked on heavily used, complex templates. the only advantage I'm seeing to your proposal is that it allows you to quickly test the template on a large variety of pages without the bother of editing those pages and changing the template name. however, you would basically expose all of wikipedia to the test version. the main reasons we shouldn't directly edit massively transcluded templates in template space is so that (a) we don't mess up thousands of pages at once, and (b) so that we don't kill the servers forcing them to re-render and re-cache all of those thousands of pages more than they absolutely need to. having this test mode would be just like editing the template directly in template space.

now, that being said, if there were some way to set up a user testing mode, so that this 'being tested' template would only appear to the particular user doing the testing (and not affect wikipedia more generally) that might be useful. I suppose we could create a mode which bypasses caches entirely for that given user and rerenders each page on the fly, using templates from an alt-template space, though I have no idea about the technical feasibility or desirability of such a thing. --Ludwigs2 19:57, 9 January 2010 (UTC)

"however, you would basically expose all of wikipedia to the test version." I don't think you understand what I am asking. If this was a special function in wikipedia, this would only be invoked by those users that requested that testing mode, set the switch, or however itthis special browsing mode would be turned on. Like setting your preferences category box is on top instead of on the bottom or changing the UTC times to be relative to the local time for this user. The wikipedia software recognizes that it will set up the display of the page a bit differently for this user, while this option is on. The users testing would turn on this template testing function, with perhaps an option to point to that users test space or a special area for this. Then for those users, the software would check this special place first for any template called for any page. If the software did not find the template in the special testing area, then it would call it in from the regular template space. It is a change in the "concatenation" of what gets called in to display a page. This would enable a group of users could just play with this for a couple of days with many pages just to make sure the template (or templates) worked correctly for a wide variety of pages. No other users would be affected. Only users turning on this option would be affected. Then when the group was satisfied, they would agree to have the template moved to the main template area and they would turn off the testing option in their preferences. stmrlbs|talk 20:24, 9 January 2010 (UTC)

That would require to bypass the parser cache, which makes it rather inefficient (doubling information stored for every page viewed). It also would have affects on Whatlinkshere and similar type of tables. If you check your preferences, do you see any option that affects page CONTENT instead of page interface (except for gadgets which work in your browser, not on the server) ? No you won't and there is a reason for that. Besides, we have things like http://test.wikipedia.org and labs.wikimedia.org to run wild with. —TheDJ (talk • contribs) 11:51, 11 January 2010 (UTC)

The BIG advantage of per template sandbox and testcases is that the testcases can be improved and include regressions, so it's not simply a matter of one template coder taking the entire test harness development burden. The good test case pages show existing vs sandbox for all the relevant configurations. If you want a more thorough test then you could make a local wiki and import all the relevant pages. Rich Farmbrough, 19:32, 11 January 2010 (UTC).

Functionality Required

In Wikipedia:When_to_use_tables#Images, I'm using <blockquote></blockquote> in this way:

<blockquote style="background: white; border: 1px solid black; padding: 1em;">
{| align="right"
| [[Image:Westminstpalace.jpg|160px|alt=A large clock tower and other buildings line a great river.|Example]]
|}
</blockquote>

What results is a white box with a black border. The picture being on the right with a white border. But I don't want this, instead, I want the whole white box to ENCLOSE the picture. Why is this not happening?

If this is a bug, can you report it to bugzilla (retorical question).174.3.101.61 (talk) 21:45, 9 January 2010 (UTC) —Preceding unsigned comment added by 174.3.101.61 (talk) 21:44, 9 January 2010 (UTC)

Another problem I have is when I use <blockquote></blockquote> like this:

:<blockquote></blockquote>

the colon (":") does not "indent" (move the whole "thing" over (to the right)). What is the reason? Why is the functionality so poor?174.3.101.61 (talk) 21:47, 9 January 2010 (UTC)

Can you report this to bugzilla?174.3.101.61 (talk) 21:49, 9 January 2010 (UTC)

I'm not sure whether it's a bug or not, but I'm also not sure why you're using a table inside a blockquote. either of the following works:

{| align="right" style="background: white; border: 1px solid black; padding: 1em;"
| [[Image:Westminstpalace.jpg|160px|alt=A large clock tower and other buildings line a great river.|Example]]
|}

<blockquote style="background: white; border: 1px solid black; padding: 1em; ">
[[Image:Westminstpalace.jpg|160px|alt=A large clock tower and other buildings line a great river.|Example]]
</blockquote>

you can indent blockquotes

::<blockquote>this is some text</blockquote>

yields:

this is some text

but it's sometimes easier just to use the {{quotation}} template:

::{{quotation|this is some text}}

yields:

this is some text

Blockquote is not strictly a MediaWiki feature. Its just an HTML tag that MediaWiki allows to be used. MediaWiki doesn't do any special processing for blockquote besides parsing the content like any other wikitext. The presentation of it is determined by the browser. Putting a table with images in it is not the intended use for blockquote, so its not incredibly unexpected that it doesn't do what you want. Mr.Z-man 03:41, 10 January 2010 (UTC)

If you need more options than {{quotation}}, then there are {{quote box}} and {{quote box2}} and more in Category:Quotation templates. ---— Gadget850 (Ed) ^talk 03:48, 10 January 2010 (UTC)

Sometimes, one would be quoting a table. That's the first/an example that comes to mind.174.3.101.61 (talk) 20:49, 10 January 2010 (UTC)

as a rule, you need to be a bit creative. wikipedia does a lot of code expansion building a page, and there's always going to be some odd case where one code expansion trips over another. If you try one thing and it doesn't work, try another. that's what the preview button is for. --Ludwigs2 22:04, 10 January 2010 (UTC)

I disagree. Wiki is supposed to be easy. I'm just saying this needs to be fix. Of course I try to fix the problem myself. But this is a problem I can't fix.174.3.101.61 (talk) 20:53, 11 January 2010 (UTC)

What links here and navboxes

When looking at "What links here" for an article, why doesn't clicking on "Hide transclusions" hide links that are only contained in common navboxes? For example, if I go to the article Exile on Main St. and click on "What links here", one of the articles listed is Got Live If You Want It! (album). If I then click on "Hide transclusions", "Got Live If You Want It! (album)" is still listed as having a link to "Exile on Main St.". But, looking at "Got Live If You Want It! (album)", the article itself doesn't contain any links to "Exile on Main St.", it's listed because it contains the {{The Rolling Stones}} navbox, which does have a link to "Exile on Main St." It seems to me that clicking on "Hide transclusions" should hide that link. So, am I missing something? Or is this a "broken" -- and if so, how can a fix be requested? (See also: Help talk:What links here#Question. How do we hide excessive linkage? — Mudwater ^(Talk) 16:30, 10 January 2010 (UTC)

Even though the link is included via a transcluded template, it is still a link and not a transclusion, so I don't think it's broken. — Martin (MSGJ · talk) 16:32, 10 January 2010 (UTC)

Right. See Help:What links here#Transclusions. "Hide transclusions" in WhatLinksHere for page A only hides pages where page A itself is transcluded and not pages where a link to page A is part of a transclusion. There is a feature request at bugzilla:3241. Articles are very rarely transcluded so "Hide transclusions" has no effect for them. It can have a large effect when WhatLinksHere is used on templates. However, many navigation templates have a "v" link to view their template page and then "Hide transclusions" loses its effect because any page transcluding the template will also link to the template page. PrimeHunter (talk) 03:00, 11 January 2010 (UTC)

At Help:What links here#Transclusions it says that What Links Here "also includes links which exist on certain pages due to the transclusion of other pages (templates). For example, if page A contains a transclusion of template B, and B contains a link to C (not contained within <noinclude> tags), then the link to C will appear on page A, and A will be listed among the backlinks of C." So, in my example, "Got Live If You Want It! (album)" is A, and contains a transclusion of template B, the Rolling Stones navbox, which contains a link to C, "Exile on Main St.". So A, "Got Live If You Want It! (Album)" is listed among the backlinks of C, "Exile on Main St.". In other words, it's a transcluded link. That's why I think that Hide Transclusions should take those off the list. So, yeah, either it's broken, or I'm still not getting it. — Mudwater ^(Talk) 04:32, 11 January 2010 (UTC)

You are still not getting it. "Hide transclusions" on page A does not hide pages where a link like [[A]] is part of a transcluded page. It only hides pages where the whole page A itself is being transcluded like with {{A}} or {{:A}}. Article pages such as Exile on Main St. are generally not being transcluded anywhere so let's take a template example. Template:OEIS is transcluded on many pages so Special:WhatLinksHere/Template:OEIS is long. But if you click "Hide transclusions" there then the list becomes limited to a few pages which link directly to Template:OEIS without transcluding it. PrimeHunter (talk) 11:26, 11 January 2010 (UTC)

Okay, I see what you mean. So, it's not broken. But, I wonder if there would be a way to enhance What Links Here so that there was an option to hide links to an article that are in transcluded objects like navboxes, like I was talking about. That would actually be pretty useful, to see a list of articles that have the actual links, with the template itself being listed, but not articles that link only through the template. Anyway, thanks for explaining. — Mudwater ^(Talk) 14:35, 11 January 2010 (UTC)

You can comment or vote at bugzilla:3241. Developers are busy and a simpler specific request may have a better chance than the suggested tree-like structure. Note that many pages are generated via a complicated interaction between source and template, for example with link generation depending on template parameters. A relatively simple sounding feature: WhatLinksHere should have an option to only list pages that would have linked if no transclusions had been performed on them. It's possible to construct examples where it gives poor results, for example when a transclusion prevents links that would otherwise have been on the page. A possibly more complicated or expensive feature: WhatLinksHere should have an option to only list pages that would link whether or not transclusions are performed on them. PrimeHunter (talk) 16:35, 11 January 2010 (UTC)

Interesting. Thanks for the info. — Mudwater ^(Talk) 22:14, 11 January 2010 (UTC)

Recent changes to the HTTPS version

I use the HTTPS version of Wikipedia to avoid my password being sent unencrypted. For the last few hours, whenever I open a page on the HTTPS version, Firefox 3.0.17 has been telling me "You have requested an encrypted page that contains some unencrypted information. Information that you see or enter on this page could easily be read by a third party."

Have there recently been changes to the HTTPS version that might account for this? (I imagine that encrypting data that is common to every page would be wasteful.) If so, does anybody know whether it is possible to disable this Security Warning for some sites and not others in Firefox? If not, does anybody have any ideas about why I might have started getting this Security Warning? If I login to the ordinary (non-HTTPS) version of Wikipedia, will my password be submitted in cleartext?

Tim Ivorson 2010-01-10

http://wikitech.wikimedia.org/view/Bits_and_pieces is causing your problem. Josh Parris 17:26, 10 January 2010 (UTC)

To be more correct.. Due to the switch to bits.wikimedia.org, MORE elements will be coming from external servers (images were already coming from upload.wikimedia.org) These elements are not encrypted and FF warns of that. I'm not sure what the plans are on this front. Will inquire. —TheDJ (talk • contribs) 18:41, 10 January 2010 (UTC)

But Firefox only started warning me today, so upload.wikimedia.org isn't causing the warning. On the other hand, I assumed it was the domain that mattered to HTTPS and upload.wikimedia.org is as much different from secure.wikimedia.org as bits.wikimedia.org is, so upload.wikimedia.org should cause the warning. I still don't get it. Tim Ivorson 2010-01-10

Thanks for the link, Josh. So part of every secure.wikimedia.org/wikipedia page I get today is coming from bits.wikimedia.org? If only there was a way of using the unencrypted Wikipedia with an encrypted password (like some Google sites). I do feel guilty about letting the Wikimedia servers unnecessarily (as far as I'm concerned) encrypt and decrypt so much stuff that's publicly available anyway. Tim Ivorson 2010-01-10

Mea culpa - I had the edit, just didn't save it - fixed now. 19:45, 10 January 2010 (UTC) —Preceding unsigned comment added by Midom (talk • contribs)

It seems to be working as before now. I wonder how you did that. Tim Ivorson 2010-01-10

We have to provide exceptions to certain changes in SSL configurations - though I did one of them, forgot to put in another. Thanks for reminder. 78.60.25.84 (talk) 22:44, 11 January 2010 (UTC)

The change was undone due to technical difficulties. The sysadmin (domas aka midom) had forgotten to save the undo of the change for the secure server however. The behavior might return when the server issue has been resolved. —TheDJ (talk • contribs) 23:25, 10 January 2010 (UTC)

Did You Know Check Script

Hi everyone

OK, earlier on I tried to install User:Shubinator/DYKcheck script check on to my monobook js file so I can check that an article I've put forward for Do You Know complies for article length. From what I can see I've installed the script correctly, I've bypassed my browsers cache but I can't seem to find out how to operate this script! Nothing new is showing up in my toolbars from what I can see. I'm completely stumped as I've followed the installation word for word. Any ideas? --5 albert square (talk) 03:01, 11 January 2010 (UTC)

You should have a DYK check link in the toolbox at the left hand side of the page (but it's only visible in article space). – ukexpat (talk) 03:09, 11 January 2010 (UTC)

Nope, definitely not getting that.

I go into the article space, all I get at the top left hand corner of the page is the page name. Under that I just get the usual navigation options and the options I have installed for Twinkle.

I go to edit the page to see if anything different appears, again all I get is my usual Twinkle options.

Have I got to enable something for this because I'm just totally not seeing what I'm supposed to get! --5 albert square (talk) 03:28, 11 January 2010 (UTC)

Well what skin are you on? In MonoBook the toolbox is in a separate column to the right of the title, underneath the Wikipedia logo. OrangeDog (τ • ε) 12:45, 11 January 2010 (UTC)

You have some errors in your monobook.js:

1. Comments that don't do anything have to start with //.
2. You have some SQL-like statements there. I don't know what they are, but they clearly aren't valid JavaScript.

I fixed them for you, the fixed version is at [20], copy it to your monobook.js. Svick (talk) 13:01, 11 January 2010 (UTC)

Hi everyone, ok I've changed to use monobook and I'm still not see this DYK script. Apparently I should be looking for something called DYK Check? The only other options I'm getting is Rules, Suggestions, Discussion, Prep Areas, Queue, Archive and Stats? --5 albert square (talk) 22:22, 11 January 2010 (UTC)

Yes, in the toolbox on left should be a link “DYK check”. Your current monobook.js works for me, did you bypass your cache? Svick (talk) 22:34, 11 January 2010 (UTC)

Hi, yes I did bypass my cache, however what I've done now is totally cleared out my cache manually and it is there now! Thanks for everyones help :) --5 albert square (talk) 00:50, 12 January 2010 (UTC)

The new editing interface

I noticed that when I click on signature nothing happens. Can the changes made to the interface yesterday be reverted until that is fixed? -- Blanchardb -^{Me•MyEars•MyMouth}- timed 14:22, 11 January 2010 (UTC)

It's working for me. And as far asI'm aware, there were no changes made yesterday (other then bits deployment and disabling again). It's more likely that a connection failed and that you have an incomplete copy of a Javascript file in your browser cache. Try WP:BYPASS —TheDJ (talk • contribs) 15:42, 11 January 2010 (UTC)

That might explain it. I had noticed a diff without green and yellow backgrounds too, which led me to believe it was a new interface. Anyway, the browser window I was using eventually froze. I had to close it through the Task Manager, and everything seems to be fine now. -- Blanchardb -^{Me•MyEars•MyMouth}- timed 16:51, 11 January 2010 (UTC)

I had a similar experience. Things have been terribly slow for me today. It appears (to me) that the issue is the bits.wikimedia.org server being unable to handle the load. Delicious carbuncle (talk) 17:32, 11 January 2010 (UTC)

Same here, though just now (literally loading this page) it seems to have cleared up. What was dissapointing is that a lot of stuff now appears to be outsourced to the bits.wikimedia domain, such that blocking it forced Wikipedia to load in text-only format with no user scripts- though the speed of loading was EXTREMELY fast. --King Öomie 17:44, 11 January 2010 (UTC)

Everything seems to be back to normal for me as well, but I don't think bits.wikimedia.org is getting accessed anymore, so perhaps a change has been backed off. Anyone know? Delicious carbuncle (talk) 21:08, 11 January 2010 (UTC)

No bits is back in business (without problems as far as I know). The problem today was that all kinds of servers (squid caches) went offline after being hammered with extra traffic (reason not yet known). If the squids go offline, the servers have a higher load then normal, and that caused some sql servers to become overly busy with traffic. Some requests were blacklisted, bits was turned on again and an extra server was put online to counter the problem. We hit 10Gb/s today btw, probably for the first time ever. Part of that traffic was because the error pages served to the blacklisted requests, were bigger pages, than the actual results of those blacklisted queries. —TheDJ (talk • contribs) 23:29, 11 January 2010 (UTC)

See also [21] [22] and [23]. Bits was turned on and off a few times in the past eight hours to find the optimal configuration for it. Still the original post was likely still caused by a broken copy in browser cache. —TheDJ (talk • contribs) 23:35, 11 January 2010 (UTC)

Something odd happening with the interface

Since yesterday—Sunday, Jan 10—I've been repeatedly getting the simple version of the interface, then it suddenly switches back to normal. Today, Jan 11, it happened right in the middle of an edit. I clicked on preview at 16:56 UTC, and the simple interface appeared. Has anyone else been experiencing this? It has been accompanied by a significant slowdown. SlimVirgin ^{TALK contribs} 17:08, 11 January 2010 (UTC)

Saw this on Sunday as well, but not since. Acroterion _(talk) 23:16, 11 January 2010 (UTC)

This happened due to the repeated on and off switching of the new bits server. —TheDJ (talk • contribs) 23:38, 11 January 2010 (UTC)

Text flowing onto a picture

The word in is in the picture. How ironic apt.

The Deluge myth article has text overflowing a picture. In my browser, the text doesn't only violate the thumbnail border of the picture, but goes into the picture itself. See screenshot to the right. This is Firefox 3.5.7 under Windows XP. Comet Tuttle (talk) 23:52, 11 January 2010 (UTC)

Hah, the irony is great. I've moved the {{commons category}} template to the bottom of the page; I'm pretty sure the number of floating elements (heh) was causing the overflowing text. EVula // talk // ☯ // 23:58, 11 January 2010 (UTC)

This is correct HTML behavior. The top of the text is higher then the top of the content of the floating element, and thus it takes higher precedence than the margin and border around the image. It usually shows when you have multiple floating elements stacked. The WP:BUNCH behavior is a problem related to this one. There is a bugzilla report on this I believe, but basically, it can only be dealt with by removing the top border/margin of our thumbed images. —TheDJ (talk • contribs) 00:05, 12 January 2010 (UTC)

Or, in some instances, of moving external link templates to the external links section. ;) EVula // talk // ☯ // 00:08, 12 January 2010 (UTC)

bugzilla:5118 and bugzilla:20640 (which i just closed as a dupe). —TheDJ (talk • contribs) 00:21, 12 January 2010 (UTC)

Curious that Wikipedia seems to be the only major website which makes such a mess of such apparently basic things. Ian Spackman (talk) 00:49, 12 January 2010 (UTC)

Wikipedia is the only major website that anyone can edit, and as result, it sometimes makes mess of such basic things as spelling, so there is no wonder it sometimes makes a mess of more complex things like floating of HTML elements. Svick (talk) 00:54, 12 January 2010 (UTC)

On any other website, floating elements with CSS is handled by a webmaster far, far away from the eyes of the average user. Wikipedia's problems are not unique, just transparent. EVula // talk // ☯ // 01:18, 12 January 2010 (UTC)

Indeed, HTML was not designed for robustness in an environment of massively collaborative editing, because nobody had thought of that application when it was designed. The wonder is that we can pull this off at all. — Gavia immer (talk) 01:20, 12 January 2010 (UTC)

Continuing (by code change) freeze of high quality animated gif File:Phenakistoscope 3g07690a.gif

(Excuse duplicate topic, but I'd already marked earlier one resolved. Want to be very specific about it being this particular high quality image which is a Featured image? on de.Wikipedia)

re animated gif: File:Phenakistoscope_3g07690a.gif

Yes, I know there's bugzilla:22041 on related things (which probably covers the problem), but should I just go ahead an try to make smaller version and upload that?

NOTE: This is a featured image on de.wikipedia and a shame it won't animate anymore. NOTE: In the upload space, it animates. -- Proofreader77 ^(interact) 00:30, 12 January 2010 (UTC)

All the information known about the issue is detailed in the bugzilla report. Making a smaller version might work, if immediacy is your target, but it is not a solution. —TheDJ (talk • contribs) 00:35, 12 January 2010 (UTC)

Yes, no solution, but I somehow suspect this kind of "bug" is LOW on the totem pole for fixing. (And, perhaps a smaller animated gif file makes more sense, anyway. I mean, really, several million megs for an animated gif? LoL Just ain't natural, is it? :-) Proofreader77 ^(interact) 02:10, 12 January 2010 (UTC)

Suggestion For The Page View Tool

I suggest that we can have the window/tool open, and when someone goes to the page, it changes\updates on real time. Report to bugzilla to fix.174.3.101.61 (talk) 04:34, 9 January 2010 (UTC)

Given that Wikipedia gets as many as 75,000 hits per second, this would probably be too resource intensive. Mr.Z-man 17:13, 9 January 2010 (UTC)

Wikipedia is the 7th most viewed site on the internet...to provide real time traffic statistics would be fairly expensive in both bandwidth and extra servers.Smallman12q (talk) 22:48, 12 January 2010 (UTC)

Google Earth

Has anyone noticed in the past few months that every time you click on a coordinate, you get the nearest street address? Kevin Rutherford (talk) 16:27, 10 January 2010 (UTC)

Sounds similar to Template talk:GeoTemplate/Archive 10#Incorrect results. — Dispenser 17:14, 11 January 2010 (UTC)

Ah, thanks for the help. Kevin Rutherford (talk) 16:34, 12 January 2010 (UTC)

Watchpages

Resolved

 – Initial user has failed to clarify discussion; widespread opposition. ╟─TreasuryTag►Captain-Regent─╢ 22:56, 12 January 2010 (UTC)

We need to be able to see the watchers of a page, and we need to see what one is watching / and people's watch lists.

Post this to bugzilla.174.3.101.61 (talk) 20:50, 11 January 2010 (UTC)

Why? Svick (talk) 21:10, 11 January 2010 (UTC)

I can't think of any reasons why this would be a good idea. Ale_Jrb^talk 22:26, 11 January 2010 (UTC)

As an editor of Wikipedia, I think my watchlist is private information of mine, and it's no business of yours what I am watching. Comet Tuttle (talk) 23:53, 11 January 2010 (UTC)

I'd like to know who has my userpage watchlisted, but I can't honestly say that I need to know. There is a tool[24] for seeing how many people are watching a page, however. EVula // talk // ☯ // 23:59, 11 January 2010 (UTC)

I'm with EVula, Comet Tuttle, and Ale_Jrb - I may have a burning curiosity to know who is watching me, but the benefit to being able to watch privately far outweighs the benefit to satisfy my curiosity. davidwr/(talk)/(contribs)/(e-mail) 00:35, 12 January 2010 (UTC)

I'm watching one of the editors in this section... and he/she/it knows who he/she/it is. bahamut0013^words_deeds 01:12, 12 January 2010 (UTC)

Bwahahah I'm doing three! ~ Amory (u • t • c) 22:53, 12 January 2010 (UTC)

< Marked "resolved" due to the peremptory nature of the request/demand, and that IP's failure to clarify their intentions later. Clearly there is a consensus that this is a Bad Thing. ╟─TreasuryTag►Captain-Regent─╢ 22:56, 12 January 2010 (UTC)

Rollbacking

OK, apologies if this is the wrong place to ask this in. I've got rollback rights, just today though I've noticed that if I go to view the difference between two page edits and decide to roll back the most recent edit, instead of taking me to the page telling me that the roll back has been successful, I'm being taken to a page about the users contributions. When I press the back button to go back to the article page and I then go into the history of the page, I can see that the roll back has been successful. If I were to compare the difference between two page edits and then go to the page history and roll back that way I do get the correct screen up.

I'm just wondering, am I the only one experiencing this issue? --5 albert square (talk) 08:56, 12 January 2010 (UTC)

There's a gadget that supposedly does that: "After rolling back an edit, automatically open the contributions of the user rolled back."
Do you have that checked in your preferences? Amalthea 09:05, 12 January 2010 (UTC)

Ah, d'you know what, I fiddled about with my preferences the other night because one of my features wasn't running at all. You're right Amalthea that was checked, I must've checked that the other night and not realised it! Now unchecked and I've just done a test edit on the rollback test page and everything is back to normal. That serves me right for fiddling with my preferences! Thanks Amalthea! --5 albert square (talk) 09:15, 12 January 2010 (UTC)

GIF thumbnail animations

A year ago all of these gif thumbs animated correctly: User:Gergyl/Glacier images Now only the first does (in Firefox 3.5 and IE8). What changed?--Gergyl (talk) 10:36, 12 January 2010 (UTC)

Whoops, sorry, I see that's already logged.--Gergyl (talk) 10:43, 12 January 2010 (UTC)

Technical problems spanning several pages

Layered images often not displaying in IE8

There is a problem with maps not displaying properly in IE that affects a number of templates. The relevant discussions are at:

• Talk:Rùm#Map_is_broken
• Template_talk:Infobox_Scottish_island#Problems_with_maps
• Template_talk:Location_map#Map_does_not_work_properly_in_IE
• Template_talk:West_Bank#Map_is_broken_in_IE
• Wikipedia:Village_pump_(technical)/Archive_69#Scottish_island_infobox_map

There may well be other problem templates that I'm unaware of.

The issue was raised here a while ago but got shunted off into the archives with no resolution. I want to make one central discussion, and direct people to it from the various pages, but I don't know where to put it. It's no good having it here if it will be archived after a few days' inactivity. What is the best way to handle this? 86.134.10.88 (talk) 23:39, 6 January 2010 (UTC).

Some experienced user with IE8 needs to log out and see if he can reproduce the problem. Since other users have already stated that they don't have this problem in IE8, I'm not sure how to find ways to reproduce the issues. —TheDJ (talk • contribs) 12:13, 7 January 2010 (UTC)

Right, but my question here is not about how to fix the problem, it is about where to put a centralised discussion of the problem, since discussions on this page seem to have a very short shelf life. 86.146.46.12 (talk) 15:17, 7 January 2010 (UTC).

The talk page of the infobox is probably the best location. —TheDJ (talk • contribs) 18:07, 7 January 2010 (UTC)

As I said, there are several different talk pages, currently each with a different discussion. I thought it would be clearer to point to a centralised discussion on a general-purpose problem-discussion page. 86.146.46.169 (talk) 18:46, 7 January 2010 (UTC).

I've had this problem a lot using IE8 and have seen it in articles transcluding each of the above templates.

This appears to be a z-index bug in IE8 but it would be helpful to know of any a workaround, preferably something that can be applied on the server side.

I've tried logging out and, to my surprise, the problem persists when browsing anonymously, so it is not caused by navpopups.

As far as I know, my IE8/WinXP configuration is not unusual.

Note that {{West Bank}} uses {{West Bank Labelled Map}} and is independent of {{Location map}}, so this is not a problem that can be traced back to any one template.

— Richardguk (talk) 21:06, 9 January 2010 (UTC)

Could this be related to wikipedia.org being on Microsoft's IE8 "compatibility view" list? Perhaps if the website were to default to standards view in IE8, the z-index image problem will go away. — Richardguk (talk) 10:51, 13 January 2010 (UTC)

Imagemap editor

I am not sure if the image map editor provided at this link http://toolserver.org/~dapete/ImageMapEdit/ImageMapEdit.html?en works or if I am using it incorrectly. When I provide the link of the picture (http://commons.wikimedia.org/wiki/File:Man_shadow_anatomy.png) in th text boxes for which I want to add imagemap and then I press the load button, the image doesn't get loaded in the image box. Please provide guidance how I should use it to add/generate imagemap code easily using this tool. Please let me know if you need more details. Thanks Amol.Gaitonde (talk) 14:52, 10 January 2010 (UTC)

It works here if I follow the instructions.

1. Load from Wikimedia project
2. Enter commons.wikimedia.org for the field "Wikimedia project"
3. Enter "Man_shadow_anatomy.png" for the field "Name"
4. Press "Load"

—TheDJ (talk • contribs) 18:30, 10 January 2010 (UTC)

Thanks for your guidance, I am able to use it now. Amol.Gaitonde (talk) 02:41, 11 January 2010 (UTC)

When I tried with the same instructions for another file http://commons.wikimedia.org/wiki/File:Circulatory_System_en_edited.svg it doesn't load the image. Even the other box in which I tried to provide the whole URL, with name of the image, it doesn't load the image. Thanks. Amol.Gaitonde (talk) 03:09, 14 January 2010 (UTC)

wiki project errors

Our wikiproject is receiving some strange text on the project page, starting around the userbox area. I think it has something to do with this: http://www.mediawiki.org/wiki/QINU_fix, but its beyond my experience. Any assistance would be appreciated. Sephiroth storm (talk) 20:16, 11 January 2010 (UTC)

It might help if you could link to the page you are talking about ;) — Martin (MSGJ · talk) 21:12, 11 January 2010 (UTC)

I believe he's talking about WikiProject Software. Svick (talk) 21:16, 11 January 2010 (UTC)

The UNIQ...QINU stuff currently on Wikipedia:WikiProject Software disappears if you remove {{Special:RecentChangesLinked/Wikipedia:Version 1.0 Editorial Team/Software articles by quality log}} or replace it with a link to Special:RecentChangesLinked/Wikipedia:Version 1.0 Editorial Team/Software articles by quality log. PrimeHunter (talk) 23:20, 11 January 2010 (UTC)

Sorry, its Wikipedia:WikiProject Computer Security. Sephiroth storm (talk) 02:08, 12 January 2010 (UTC)

Okay, I think PrimeHunter's solution worked. Sephiroth storm (talk) 02:14, 12 January 2010 (UTC)

Sorry for barging in as I was working on this problem in WIkiProject SOftware. Is there anyway you can fix this problem so that we can transclude the page without the weird error codes? --Tyw7  (Talk • Contributions)   Changing the world one edit at a time! 16:06, 13 January 2010 (UTC)

bits and prototype servers

Is there a continuing problem with these servers? At the moment everything seems to be very sluggish and it appears to be stuff being served from these servers that's the problem. Thanks.-ukexpat (talk) 19:02, 12 January 2010 (UTC)

http://wikitech.wikimedia.org/view/Server_admin_log So yes, prototype was having trouble. There have been no reports of bits failing, though there has been some unreachable problems in amsterdam today. —TheDJ (talk • contribs) 19:43, 12 January 2010 (UTC)

OK thanks. Things seem to have improved in the last 30 minutes or so. – ukexpat (talk) 19:54, 12 January 2010 (UTC)

Uh oh, I am only getting the simple text UI at the moment. Do we have a problem, Houston?-ukexpat (talk) 01:02, 13 January 2010 (UTC)

Nope. The Wikipedia style sheet seems drastically altered. Or maybe it's just been reselected for me? Interesting, the page content now references http://bits.wikimedia.org/w/extensions/UsabilityInitiative/css/combined.min.css?16 and http://bits.wikimedia.org/w/extensions/UsabilityInitiative/css/combined.min.css?16 I'm pretty sure those weren't there before. —EncMstr (talk) 01:10, 13 January 2010 (UTC)

The reference is because bits now serves the css and js files, where it didn't before. Ale_Jrb^talk 08:22, 13 January 2010 (UTC)

Rendering issue?

Resolved

Did Wikipedia start rendering all funky for anyone else all of a sudden? –xeno^talk 01:08, 13 January 2010 (UTC)

Yup, see above - I am getting the simple text UI.-ukexpat (talk) 01:10, 13 January 2010 (UTC)

Yar should have used a heading that dummies could recognize =) –xeno^talk 01:13, 13 January 2010 (UTC)

No, not for me. --DThomsen8 (talk) 01:11, 13 January 2010 (UTC)

For me it has for my logged-in account through Safari. Logged out in a Firefox window is fine. maybe a CSS issue with Vector? Safari's status bar is telling me 13 errors on the Main Page. Stu 01:12, 13 January 2010 (UTC) —Preceding unsigned comment added by Stu (talk • contribs)

I just logged-in with Firefox and am experiencing same problem. So for me it's just when logged-in. I use Vector.Stu 01:14, 13 January 2010 (UTC) —Preceding unsigned comment added by Stu (talk • contribs)

Everything but my watch page renders. Bidgee (talk) 01:16, 13 January 2010 (UTC)

Known issue now, they are looking for the guy to fix it. (bits is partly down again apparently) —TheDJ (talk • contribs) 01:17, 13 January 2010 (UTC)

Same here, also using Safari. Most or all resources located on bits.wikimedia.org have a "lost network connection" error and are not being fetched. It's happening on Commons and Wikisource too, when logged in. Ah, OK per the above. Carl Lindberg (talk) 01:18, 13 January 2010 (UTC)

Seems to be back working now. IRC discussion at #wikimedia-tech ongoing.Stu 01:25, 13 January 2010 (UTC) —Preceding unsigned comment added by Stu (talk • contribs)

Bleh. Nothing is working for me while logged in in either Chrome or Firefox. user:bkonrad 01:33, 13 January 2010 (UTC)

I'm still having the problem both logged in and out, in Firefox 3 and IE 6. Useight (talk) 02:05, 13 January 2010 (UTC)

Everything ok on my side now. –xeno^talk 14:50, 13 January 2010 (UTC)

RecentChangesLinked: duration more than 30 days?

Resolved

 – it's a bug or a feature

Is it possible to check Related changes for a duration of more than 30 days? I changed the "days" parameter in the URL for several pages, such as this one http://en.wikipedia.org/w/index.php?title=Special:RecentChangesLinked&limit=250&days=100&target=WIMPs, but I never get beyond 30 days. The only change - as if to spite me - is that the page now displays a bold "100 days" option. This has been moved here from WP:HD. — Sebastian 02:17, 13 January 2010 (UTC)

Theoretically possible, but would require a change in the MediaWiki config, which is currently set to 30 days for all Wikimedia projects ($wgRCMaxAge = 30*86400;). Unless there's an extremely good reason to increase this, I can't see this being done. Amalthea 10:30, 13 January 2010 (UTC)

Thanks for your answer. This seems like a very arbitrary limitation; what is the reason for it being 30 days, and why do you say it needs an "extremely" good reason to be increased? If it has to be that way, can it at least be changed to react reasonably, instead of displaying the misleading option "100 days"? (I know this is not the place for feature requests and bug reports; but since the conversation went that way, I thought it couldn't hurt to ask.)

The reason I'm asking is probably not "extremely" good, but I can't think of an alternative: A WikiProject with some 30 project related pages, which used to be very active, is currently in hibernation, with periods of inactivity that easily exceed 30 days. The watchlists we keep at WP:SLR#Watchlists provide a way to ensure that we don't miss any edits, be they in good faith or malicious. Is there a better way to do that? — Sebastian 18:00, 13 January 2010 (UTC)

What non-arbitrary number of days do you have in mind? :) Special:RecentChanges doesn't get its data from the normal page histories, but from the "recentchanges" database table that was designed for this job, mostly for performance reasons; I assume performance is also the reason for keeping the entries for a limited time only, but there might also be privacy concerns since user IP addresses of logged-in users are stored there as well. When I say you need an "extremely" good reason I'm mostly speaking from experience over at bugzilla, where you'd have to request the change: I don't see that they'll change it without a very good reason. But feel of course free to make the request—just telling you what to expect. :)
Considering your concern of missing edits, maybe the upcoming patrolled revisions can help you with that? I read they are planning a mid-January release (sounds a tad optimistic to me though). If not, you could ask someone to build a toolserver tool which, given a list of pages, returns a list of recent changes built from the page histories, which could then in fact be of arbitrary length. Doesn't sound hard to do, I'm sure you could find someone to help you out.
Cheers, Amalthea 19:04, 13 January 2010 (UTC)

Thanks, I will consider that. This resolves my question, the purpose of which was only to find out if there was a limit. — Sebastian 21:43, 13 January 2010 (UTC)

Footnotes and html small tags

This VPP thread is veering off into technical territory it seems, does someone here have any advice? (pls centralise comments there) MickMacNee (talk) 21:29, 13 January 2010 (UTC)

No highlighted changes when comparing diffs

For some reason I no longer have the changes highlighted when I compare new edits to the previous version. Is this just my browser, some funky account setting or a Wikipedia issue? Agne^Cheese/^Wine 01:26, 14 January 2010 (UTC)

The highlighting also disappeared for me but is back now. I use Firefox and didn't change browser or settings before or during the incident. PrimeHunter (talk) 01:55, 14 January 2010 (UTC)

Yes it is back for me too. I guess it must have been a Wikipedia thing. Agne^Cheese/^Wine 02:25, 14 January 2010 (UTC)

Pictures

Resolved

 – Smallman12q (talk) 19:15, 14 January 2010 (UTC)

Why is it that when wikimedia resamples a picture, the new size is often larger than the actual picture. For example, my File:Incarceration_rate_of_inmates_incarcerated_under_state_and_federal_jurisdiction_per_100,000_population_1925-2008.png is 29kb, but after wikimedia resamples it, it shows as 85kb? Wouldn't it save bandwidth to just send the full image rather than resample it if the resampling is bigger?Smallman12q (talk) 16:59, 11 January 2010 (UTC)

When I examine that file page, underneath the picture is this text: "Full resolution‎ (2,882 × 1,674 pixels, file size: 29 KB, MIME type: image/png)". Doesn't that indicate it's 29K? Comet Tuttle (talk) 23:55, 11 January 2010 (UTC)

Ah, yes the actual original uploaded picture is 29kb. However, wikimedia resamples it. If you right-click on the picture and select properties, you will see that its far more than 29kb.Smallman12q (talk) 00:26, 12 January 2010 (UTC)

(edit conflict) Well, yes the original image (that's 2,882 × 1,674 px) has 29 KB, the resampled version (800 × 465 px in this case) has 85 KB. It's because the original image contains only 22 colors, but the resampled has 1850 colors due to anti-aliasing. The T3218 that dealt with this issue was marked as WONTFIX, mostly because MediaWiki's resizing algorithm is presumably better than browsers'. Svick (talk) 00:42, 12 January 2010 (UTC)

Ah...I see. But now the question is better than which browser? I do believe that firefox does a better job of rescaling than mediawiki. Or am I wrong? Smallman12q (talk) 22:29, 12 January 2010 (UTC)

Firefox's is probably the worst of the ones I tested. [25] has a comparison, with the image scaled to 800x465, the top is the full image, scaled by Firefox, the bottom is MediaWiki's (using ImageMagick). The test page I used for it is here. Mr.Z-man 23:37, 12 January 2010 (UTC)

I see...thanks for clearing that up z-man. ImageMagick still does a better job than firefox.=PSmallman12q (talk) 19:15, 14 January 2010 (UTC)

Wide margin in article

Does anybody see the left margin in the Utah Beach article as being a bit wider than the norm? I traced the change to this diff, something to do with "fix bunching". I'm using IE7 in Vista. Abductive (reasoning) 00:11, 13 January 2010 (UTC)

Yeah, that edit caused it. I fixed the issue by correctly closing the {{fixbunching}} template. Svick (talk) 00:22, 13 January 2010 (UTC)

also note that the fixbunching creates double margin/padding (1 from the fixbunch box, 1 from the elements contained within it). You need to use {{FixBunching|beg|margin=0|padding=0}}. As to why that is not the default behavior, I'm guessing historic reasons, but perhaps it's time to invert that behavior. —TheDJ (talk • contribs) 00:29, 13 January 2010 (UTC)

Yes, please do. Thanks. Abductive (reasoning) 00:32, 13 January 2010 (UTC)

I have reversed the default behaviour of fixbunching into not adding extra margin. A sample of the current uses showed that this was what people were expecting. —TheDJ (talk • contribs) 19:17, 14 January 2010 (UTC)

Page notice redlink when editing a User's Talk page

When I edit a User's Talk page, there is a small redlink in the upper right hand corner of the edit page labeled "Page notice". If I click on this red link, it takes me to a link called "User talk:xxx/Editnotice" where "xxx" is the Username. What is this page, what is its purpose, and where is it documented? Woogee (talk) 02:15, 13 January 2010 (UTC)

It's to create an WP:EDITNOTICE. That page describes them, but you can see a lot of discussion at WT:EDN#Navbar. ~ Amory (u • t • c) 02:21, 13 January 2010 (UTC)

Er, uh, okay. Woogee (talk) 02:26, 13 January 2010 (UTC)

Why was this added? This is just going to lead to 1) "test page" type edit notices, 2) vandalism edit notices, 3) malplaced talk page messages. Why not hide it with CSS and have folks who want to see these opt-in? –xeno^talk 14:51, 13 January 2010 (UTC)

Never mind, I see it was added after a fairly low-trafficked conversation at Wikipedia talk:Editnotice#Navbar (which I was the first to reply, suggesting it be hidden by default...!). Suggested again at Wikipedia talk:Editnotice#User page links. –xeno^talk 15:11, 13 January 2010 (UTC)

We have been discussing to add those links for 1 year and 3 months now. The "Navbar" discussion is just the latest of many such discussions both on that talk page and here on the Village pump. We just haven't gotten around to add it until now.

And today we updated the editnotice system so you only see the red "Page notice" link on your own user and user talk page. You should not see it anywhere else. (Unless you are an admin or accountcreator, they see those links on all pages, since they can create and edit editnotices for any page.) But everyone see the blue "Group notice" and "Page notice" links on any page if the editnotices already exist.

The reason we show those links is to make the editnotice system user-friendly, since people had lots of problems of figuring out the right page names to use for the editnotices. And even for us who know the details of the system it was messy to reach the editnotices of a page.

--David Göthberg (talk) 21:59, 14 January 2010 (UTC)

 Works for me thanks for the explanation. Any CSS to hide the links for admins who don't want to see them? –xeno^talk 22:00, 14 January 2010 (UTC)

.editnotice-link {display:none !important;}

should hide them everywhere. Algebraist 22:08, 14 January 2010 (UTC)

Thanks =) –xeno^talk 22:10, 14 January 2010 (UTC)

The CSS shown above hides both the blue and the red links. It's on my to-do list to add a "editnotice-redlink" class that can hide only the red links. It will be deployed in an hour or so.

--David Göthberg (talk) 22:57, 14 January 2010 (UTC)

Y Done - And here's the code to use if you don't want to see the red "Group notice" and "Page notice links:

/* Hide the red "Group notice" and "Page notice" links. */
.editnotice-area .editnotice-redlink { 
  display: none !important;
}

Add the above code to your personal /monobook.css, then wait a minute, then bypass your browser cache. (If you use another skin, here's a link to your skin CSS.)

--David Göthberg (talk) 23:41, 14 January 2010 (UTC)

Question re performance impact of a template

Here, on another WP talk page, I offered a suggestion for avoiding the {{#Tag:Ref|...}} idiom used for nesting refs, which can look jarring to editors not familiar with it, by wrapping that in a template like (this is a bit different that the original -- I've added the group parameter)

{{#if:{{{1|}}}<!--
 -->|{{#if:{{{name|}}}<!--
  -->|{{#Tag:Ref|{{{1}}}|name={{{name}}}|group={{{group|}}}}}<!--
  -->|{{error|Error:''name'' parameter must be defined.}}<!--
 -->}}<!--
 -->|{{Error|Error:empty ref body.}}<!--
-->}}

Such a template would seem to be a drop-in nestable replacement for the non-nestable <Ref>...</Ref> except that it requires the name parameter to be supplied. It works just fine where there is no nesting, though, and I'm wondering about what the performance impact might be if such a thing came into widespread use as alternative to <Ref>. Any thoughts? Wtmitchell (talk) (earlier Boracay Bill) 07:39, 13 January 2010 (UTC)

Coding true nested refs has been on my todo list for a while, so I'd rather discourage getting too caught up in this kind of hack. Dragons flight (talk) 00:41, 15 January 2010 (UTC)

I'd be less concerned about efficiency and more concerned about introducing more inconsistency into something that's already complicated. Mr.Z-man 02:42, 15 January 2010 (UTC)

CSS screen media type in Opera's fullscreen (thread revival)

I reported a bug in Wikipedia here which I have a few questions about.

The issue was that Wikipedia's layout was (and still is) completely broken in Opera's fullscreen mode. This is because Opera acts as "powerpoint-esque" presentation software and adheres to the W3C CSS spec on presentation media-types, a media-type Wikipedia actively ignores.

Firstly, the bug is ONLY present in Wikipedia (that I'm aware of - I've tested every other WikiMedia project and none have this issue), yet it was treated (when I reported it) as a MediaWiki bug, ticketed and "fixed" (this "fix" has evidently not been pushed to Wikipedia yet though as the issue still remains).

However, the nature of the eventual "fix" and the general prevailing view that this is somehow an Opera bug or deficiency that needs to be "hacked around" is what I'm curious about - and by extension, I think the whole thing seems quite short-sighted. This particular issue just so happens to affect the Opera browser, but could potentially affect any user-agent that uses any one of the media-types Wikipedia ACTIVELY ignores. I use the word "actively" because it's not a matter of failing to include them - the default behaviour of CSS is to automatically provide styles to ALL media-types, Wikipedia takes an extra measure by adding the screen specific media attribute to ensure nothing but "screen" media types get the proper styles - a measure which makes little or no sense in my mind.

Thoughts? Am I missing something obvious here? _ɹəəpıɔnı 09:19, 13 January 2010 (UTC)

Found the cause, we had a local change on the English Wikipedia that was causing this. I have corrected this. —TheDJ (talk • contribs) 11:57, 13 January 2010 (UTC)

Thanks for that, but I think you missed the point I was making. Why does Wikipedia deliberately exclude all unconsidered media-types? No other website that I'm aware of does this.

Projection is a valid media-type just as any other - it shouldn't be necessary to avoid it's usage, nor the usage of any other media-type for that matter - this hackish approach to fixing this could just as easily lead to the same issue occurring in another non-screen user agent. I'm thankful that it's fixed, but it just seems woefully unsustainable to me. _ɹəəpıɔnı 12:38, 13 January 2010 (UTC)

Other media types generally represent other devices, that may or may not be able to handle some of the complex CSS that we use. By putting them, essentially, into a compatibility mode, we can ensure that it will work in a usable manner on as many things as possible. That's how I see it, anyhow. Ale_Jrb^talk 12:41, 13 January 2010 (UTC)

Correct, it may not be pretty, but at least it will work. Besides, it's all irrelevant. As long as screenreaders don't even support speech/aural as a media type, i doubt braille/embossed/projection/tty/tv are going to be used any time soon. It is a shame, but the software to support it, just isn't out there atm. When it does, we can easily add new mediatypes to mediawiki I think. —TheDJ (talk • contribs) 12:48, 13 January 2010 (UTC)

(edit conflict) By compatibility mode, I mean that devices rendering one of the unsupported media types should only display those stylesheets that are for all devices, which are far simpler than the skin sheets (which in turn are only suitable for a screen). In the Opera case, because Opera decides to render everything under projection when in full screen if it sees a projection style anywhere, it wasn't rendering any of the screen styles - despite being on a screen. This is Opera's fault, and imo, a bad design decision. Either way, it isn't a question of being 'short sighted' - because support can be added at any time. If you want to style everything correctly for devices that use prentation (which is probably only full-screen Opera), for example, I'm sure that nobody is stopping you. There are possibly more important things though. Ale_Jrb^talk 12:53, 13 January 2010 (UTC)

@Ale_jrb:

"Other media types generally represent other devices, that may or may not be able to handle some of the complex CSS that we use" - I'm not sure what your experience with stylesheets is, but this is not really how CSS works. IE for example is not "able to handle some of the complex CSS" that some newer sites use - yet serving such to IE is not a problem due to something called "graceful degradation". This allows you to serve any CSS to any software on any device without issue, regardless of that software's level of support. So this is a non-issue.

"By compatibility mode, I mean that devices rendering one of the unsupported media types should only display those stylesheets that are for all devices" - Again, this is not how CSS is designed. Devices not supporting any one of the "complex CSS that we use" will be automatically compatible regardless of what or how many stylesheets they are fed as unrecognised styles are ignored by CSS parsers.

As I said above, a point you have both ignored, Wikipedia is absoutely unique in this behaviour. Almost every other site on the web does as I am proposing without running into any of the issues you seem to think might occur. Even every other WikiMedia project does as I'm proposing without encountering any problems. _ɹəəpıɔnı 13:12, 13 January 2010 (UTC)

We are unique in many things. Besides, if you look at the source, you'll note that we are clearly using a mix of media settings for the stylesheets. Also graceful degradation often isn't that graceful at all. I'm quite sure Firefox 1.5 and Netscape 7 crash when visiting the English Wikipedia. (If anyone has those browsers and the skills to investigate, please do). —TheDJ (talk • contribs) 13:20, 13 January 2010 (UTC)

As a web developer, I do like to pretend that I know a little bit about CSS. Graceful degredation? Lol. Sending a simple stylesheet (i.e. only those media styles that are happy to exist on any device, such as what happened to Opera) is significantly better than a mess of half working monobook that you would get if you leave the monobook styles to degrade onto lesser browsers - have you seen it on IE5 recently? (NB. I haven't. I saw it a while ago. I haven't looked again, and never want to agai). While you're correct in that it won't technically break, by sending simple stylesheets that all devices will understand, you can get the correct intended behaviour. We do degrade - by sending less complex CSS in the event that the complicated CSS is unlikely to be supported (such as when the browser is requesting tty). The reason that some other sites look so terrible in some mobile browsers, for example, while Wikipedia looks as expected (the view you were getting in Opera) is because we do this. Ale_Jrb^talk 13:41, 13 January 2010 (UTC)

@TheDJ

Yes Wikipedia is unique in many things - often in very positive, progressive ways. Often.

@Ale_jrb

Don't worry, I am all too aware that "graceful degradation" more often than not involves more degradation than grace, but it is the terminology used. CSS parsers ignore styles they don't recognise - even in IE5. Apologies if I underestimated your knowledge of CSS, but your initial comments seemed to imply that you didn't understand this concept. You said certain software "not be able to handle some of the complex CSS" (your exact words). That is simply untrue - they "handle it" by ignoring it.

Anyway, I can see this is going nowhere. I shall wish you both well in maintaining the stylesheets across an ever increasing number of applications/platforms. Good luck. Thanks for your interest. _ɹəəpıɔnı 15:44, 13 January 2010 (UTC)

Yes, but when certain styles are ignored, it can destroy all readability of the page. Monobook (for example) is so complex, that a device that has only limited parsing of CSS will create complete rubbish when it ignores half of it, and tries to parse the other half. This is what I meant when I say that simple devidces can't handle it. When you send a very simple sheet, however, you can safely expect most devices to output it as expected. It won't look as nice, but it won't be unreable either. Your final comment about maintaining the stylesheets is strange, as you started off by telling us to add stylesheets for the unsupported media types. Doesn't this make maintaining it harder?

Either way, thanks for reporting the Opera bug. :) Ale_Jrb^talk 15:50, 13 January 2010 (UTC)

First of all, this issue is not totally unique to Wikipedia. Just Google "Opera projection CSS bug" or something. When I was researching this, I did see some others who had to hack around the problem (usually by providing media="screen,projection" on everything).

Second of all, we don't do media="all" on all our stylesheets because we want different styling for different media. Currently we provide styles for screen, print, and handheld. What do you propose we do instead? Replace every instance of media="screen" with media="braille,embossed,projection,screen,speech,tty,tv"?

The fact of the matter is that we have no stylesheets for projection as defined by the CSS2.1 standard. We don't test in projection UAs and we don't intend Wikipedia to be viewed that way. (Note that projection is a paged media type – Opera using it in fullscreen mode is IMO wrong.) UAs that are projection-only are best served by being told that we have no projection stylesheets, only screen stylesheets. They can then try to heuristically make a good projection-based style from our screen styles if they like. This is in fact what Opera does, unless we incorrectly identify screen stylesheets as projection.

I do think Opera is acting incorrectly here. It should not treat fullscreen mode as projection – that breaks user expectations. Fullscreen is still screen. How it should detect projection mode instead isn't entirely clear, though, to be fair to them. —Simetrical (talk • contribs) 18:11, 14 January 2010 (UTC)

Animated gif bug?

moved from WP:AN by  Skomorokh  17:44, 13 January 2010 (UTC)

It appears that lately, any page that has several animated gifs, have a problem - only one is in motion. Please check Animation for an example. In addition to the bouncing ball, both the horse and the earth should be moving. (I've noticed this with both Firefox and Safari... still worked a few weeks ago.) Edit: Even the single animated gif on Cartoon doesn't move. If I've posted this in the wrong forum, please inform me, or transfer the info, thanks! Greetings, --Janke | Talk 15:00, 13 January 2010 (UTC)

Probably the wrong noticeboard, but can you see two of the signatures at User_talk:LustyRoars flashing (it uses font style="text-decoration:blink" ). Not everyone seems to be able to see the words flash - possible related problem? (user's regular sig doesn't flash btw, this was just a test) --Elen of the Roads (talk) 16:04, 13 January 2010 (UTC)

I can confirm the problem described by Janke. How strange. SPLETTE :] ^{How's my driving?} 17:22, 13 January 2010 (UTC)

A lot of people complained about the <blink> tag because many people abused it (remember those old Geocities pages with bright flashing colors?), so some browsers started ignoring it. Idem for the blink text decoration. --Enric Naval (talk) 12:17, 14 January 2010 (UTC)

Flood fill has five animated GIFs, of which four work okay and one (File:Smiley fill.gif) which is frozen. I rather suspect the thumbnailing to be the issue. -- Finlay McWalter • Talk 17:49, 13 January 2010 (UTC)

Yeah, see User:Finlay_McWalter/sandbox - the same image works when full size but is frozen when thumbnailed. Images are thumbnailed by Mediawiki, so I expect something has broken recently. -- Finlay McWalter • Talk 17:55, 13 January 2010 (UTC)

#GIF thumbnail animations above says this is a known bug. -- Finlay McWalter • Talk 17:57, 13 January 2010 (UTC)

Werdna just disabled GIF scaling again, because no one who understands that code has time to fix it right now. The animated GIFs should work now. —TheDJ (talk • contribs) 18:32, 13 January 2010 (UTC)

Yes, they do, thanks. If others stil have problems, add ?action=purge to the page URL. --Janke | Talk 21:20, 13 January 2010 (UTC)

Reference Regex

I thought that some here might appreciate it if I captured this. I created a (.NET) Regex which captures inline Cite.php references:

(?<openStart>\<ref\s*)(?<refName>(group\s*=\s*|name\s*=\s*)[^\>] ){0,2}(?<close>\s*(?(?=/\>)/\>|\>\s*(?<content>[^\<]*)(?<contentClose>\<\s*/\s*ref\s*\>)))

So far I haven't come across an issue with it missing anything, or over/under capturing anything. If anyone does see any issues with it, or potential improvements, I'd appreciate a heads up. Enjoy!
— V = I * R (talk to Ohms law) 05:37, 14 January 2010 (UTC)

Ref templates

You should look at {{r}}, {{sfn}} and {{sfnRef}}. ---— Gadget850 (Ed) ^talk 12:37, 14 January 2010 (UTC)

Good point... I don't think that I'll try squeezing that into a single regex, though. It's my program after all, may as well use multiples. Thanks for pointing that out, though!
— V = I * R (talk to Ohms law) 19:20, 14 January 2010 (UTC)

Tags in ref

I'm pretty sure this breaks if you have HTML tags inside the reference. I typically use <ref[^>]*>((?:[^<]|<(?!/?ref))*)</ref>. I think we have some more complete examples in the AWB manual. — Dispenser 17:08, 14 January 2010 (UTC)

humm...I'm honestly not sure what you mean by "HTML tags inside the reference". I know that it catches url's inside of the content portion between the ref tags. Do you have a wikitext example? Thanks!
— V = I * R (talk to Ohms law) 19:24, 14 January 2010 (UTC)

I don't keep examples around but something like: <ref>Since O<sub>2</sub>'s partial pressure is the fraction of O<sub>2</sub> times [...] </ref>, other examples could include <cite>, <abbr>, or some other wiki tags like <math> or <nowiki>. — Dispenser 19:49, 14 January 2010 (UTC)

Ahh, I see. I'll test some things like that out. I think that it's OK, because I'm looking for either "/>" or the "</ref>" to end the capture, but I definitely appreciate the question.
— V = I * R (talk to Ohms law) 20:09, 14 January 2010 (UTC)

I see what Amalthea is getting at below however, regarding the content group stopping at the first bracket. Working on a change for that now.
— V = I * R (talk to Ohms law) 23:01, 14 January 2010 (UTC)

Whitespace

With the content group only matching [^<]* it'd still stop at the first angle bracket, and won't be able to match such a ref at all. And you're missing some potential whitespace issues at the opening tag, like with "<  ref name  =foo>". Amalthea 21:14, 14 January 2010 (UTC)

re: missing some potential whitespace issues, is that even a valid tag? I mean, does cite.php recognize "<  ref name  =foo>" as a reference? I could easily put "\s*" in more places, but I didn't think that whitespace was allowed between the bracket and the keyword (the code docs at mw:Manual:Code seems to suggest that at least, unless I'm missing something). Now that I think about it though, whitespace is definately allowed around the "=", so I went ahead and fixed that.
— V = I * R (talk to Ohms law) 21:39, 14 January 2010 (UTC)

Hmm, < ref>foo</ref> ... nope, you're right. I consider that a bug in Cite.php, but there you go. :) Amalthea 00:27, 15 January 2010 (UTC)

Has nothing to do with Cite, none of the tag like entities allow that kind of whitespace. Dragons flight (talk) 00:37, 15 January 2010 (UTC)

I suppose that a discussion could occur about making the tags less sensitive to whitespace, in the MediaWiki software, so that spacing like that would be valid... my concern with that tough, is that it can generally make wikiText non-XML compliant. I mean, you can't stick spaces like that into HTML tags, or any other XML tags, so... I'd suspect that adding that level of forgiveness would be a royal PITA to properly develop, as well. The only benefit that I could think of is a slightly more user-friendly wikitext. But then, I think an argument could me made that it'd just lead to lazy wikitext markup habits, especially considering the fact that the learning curve seems to be low in figuring out that <tag> is a proper tag, while < tag > isn't.
— V = I * R (talk to Ohms law) 00:54, 15 January 2010 (UTC)

wikEd

FYI, the next version of wikEd will contain a new and real wikicode parser, regexp are always poor substitutes for such complex tasks. For some tricky short code blocks to find the corresponding opening and closing html elements see the User:Cacycle/wikEd.js subroutine WikEdRemoveHighlighting. Cacycle (talk) 07:45, January 15, 2010 (UTC)

heya Cacycle, I love using wikEd, by the way. You're certainly right about needing more then regex. using one here makes the most sense for my particular use right now, though. I know that there are many who use AWB and other tools, which can also use regex's, who could use a good expression that captures all reference instances.
— V = I * R (talk to Ohms law) 22:58, 14 January 2010 (UTC)

Hovering Cat bar obscuring bottom of pages

Apologies if this is an FAQ.

Lately, I've found that when using IE v 6.0 when I scroll to the bottom of various pages, the Category bar obscures the last lines of text by hovering over them. And sometimes it doesn't. Any ideas why? Is it something in my monobook?

And no, IE v 6.0 is not my choice of browser. If I'm using it, it's because I can't use a different one. --Dweller (talk) 11:38, 14 January 2010 (UTC)

Update - it did it just now on this page, too. --Dweller (talk) 11:39, 14 January 2010 (UTC)

Have you tried disabling gadgets and blanking /monobook.js ? —TheDJ (talk • contribs) 11:46, 14 January 2010 (UTC)

No. I've not touched either of them for probably at least a year and this has just started happening. Or not happening. It seems quite random. --Dweller (talk) 11:49, 14 January 2010 (UTC)

People make changes to the scripts you include ... —TheDJ (talk • contribs) 12:38, 14 January 2010 (UTC)

This happens whereever there is a collapsible table on the page. I don't know what causes this behaviour, but I do know is that it started happening two/three months ago . — Edokter • Talk • 23:28, 14 January 2010 (UTC)

Harry Potter navbox

I use classic skin, not monobook. The Harry Potter navigation box at the bottom of the many HP pages is fine when hidden, but when shown, it grows to a width wider than many browsers, especially if not full screen. I have left a message on the HP portal talk page, but if anyone here knows how to fix it I would be grateful - it is too complex a template for me to tackle. -- SGBailey (talk) 12:47, 14 January 2010 (UTC)

browser version ? —TheDJ (talk • contribs) 12:53, 14 January 2010 (UTC)

Ah the template has a minimum width of around 970px due to the top row with the titles of the books. —TheDJ (talk • contribs) 12:58, 14 January 2010 (UTC)

Yeah, it runs slightly outside my screen too on the monobook skin (sidebar included).--Unionhawk ^{Talk E-mail} 12:59, 14 January 2010 (UTC)

If it is going to have a fixed width in pixels, I suggest 640 for a full screen VGA monitor. It would be better if it were dynamic though. -- SGBailey (talk) 15:05, 14 January 2010 (UTC)

Well it doesn't have fixed width, it really has a minimum width. This is caused because that first line is a table inside a div inside a tablecell. Tables take as much space as needed to render their content, I'm not really sure how to fix this without totally changing the design of the template. The HP guys should discuss this. —TheDJ (talk • contribs) 15:38, 14 January 2010 (UTC)

Twinkle Lost extra tabs

I used to have extra tabs for request protection, XfD and the like at the top of pages. They have now disappeared. Anyone know why/how? DuncanHill (talk) 17:58, 10 January 2010 (UTC)

I've checked my gadgets, and the box for Twinkle is still checked. DuncanHill (talk) 18:05, 10 January 2010 (UTC)

Twinkle relies on your browser's Javascript support, so if you switched browsers, enabled Javascript blocking, or got a Javascript-blocking browser add-on (like noscript), that could be the reason. Also try a hard refresh (usually shift-F5 ctrl-F5), or closing and restarting your browser. Equazcion ^(talk) 18:06, 10 Jan 2010 (UTC)

Haven't switched browsers, and Java working fine on other sites. DuncanHill (talk) 18:08, 10 January 2010 (UTC)

Try logging out and back in, and restarting your browser completely. I've had this happen before, and when it wasn't a javascript blocking issue, it was generally just a momentary browser glitch that cleared itself up eventually. Equazcion ^(talk) 18:14, 10 Jan 2010 (UTC)

Also, if you've switched to the Vector skin (or the "Try Beta" option that uses Vector), items that were tabbed under Monobook are now mostly found under the drop-down menu under the downward-facing arrow. — Gavia immer (talk) 19:03, 10 January 2010 (UTC)

Thanks for the suggestions, tried closing and opening browser, and don't use the Vector skin. Still not working. DuncanHill (talk) 22:17, 10 January 2010 (UTC)

You can also try clearing out your monobook.js page, then refreshing. It's possible one of your imported scripts has a recently-introduced bug that's messing up other scripts (like Twinkle). Equazcion ^(talk) 22:27, 10 Jan 2010 (UTC)

What browser are you using, and can you have a look into your browser's javascript console and look for any errors (Ctrl Shift J in Firefox)? Amalthea 14:20, 11 January 2010 (UTC)

Using IE8. Where would I find a Javascript console? DuncanHill (talk) 16:07, 12 January 2010 (UTC)

Using IE you don't need to, since Twinkle never has and never will work in Internet Explorer. I thought that was well documented, it says so in the Gadgets section, in the script description, and in the troubleshooting section of the documentation. Amalthea 16:15, 12 January 2010 (UTC)

You should make it so that it pops up an alert on every page load if you're using IE, saying that the rest of it won't work. It would be so worth it. :D Ale_Jrb^talk 16:19, 12 January 2010 (UTC)

Ask AzaToth about that but he'll get his mind really boggled revising Twinkle's javascripts. --macbookair3140 (talk) 15:06, 16 January 2010 (UTC)

Linked images

I didn't really know where to go with this one. Does anyone know if and how to make link to penalty card#yellow card, in a similar way to how links to Angola? Thanks in advance, WFCforLife (talk) 00:26, 13 January 2010 (UTC)

It's called the link= option, but this card cannot use it, because it won't link to the imagedescription page anymore, which in this case is required due to attribution requirements on the image. You will need a public domain image or something. —TheDJ (talk • contribs) 00:34, 13 January 2010 (UTC)

Thanks for the help. A potential solution is to go and create one myself from scratch, give it an alternative name and release it into the public domain. Out of curiosity, if I attribute on first use, does that meet the requirements of the licence? For the purposes for which I intend to use it, attributing on the first occurance would still work very nicely. WFCforLife (talk) 01:28, 13 January 2010 (UTC)

Does it actually need to hyperlink to the image page for attribution? Inspection of the source would lead you there. OrangeDog (τ • ε) 14:02, 13 January 2010 (UTC)

CC licenses are vague on that. They state that attribution must be done in a way appropriate for the medium. I think people having to look at the HTML source is looking for the limits of "appropriate" here. —TheDJ (talk • contribs) 12:40, 14 January 2010 (UTC)

Certainly you can make this link with link= (). I think that the overblown copyright claims need to be nipped right away. The image has been licensed directly to Wikipedia under the GFDL, the CCSA 3.0, the CCSA 1.0, 2.0, and 2.5 generic ... and people are telling you that you still can't a click on the image lead to the link you want? Why don't they demand that you include the full name and address of the author directly under the image each time it appears on the page while they're at it? The fact is, Wikipedia is an encyclopedia anyone can edit, and anyone who does edit the page will see where the image's home page is with all the historical and licensing details. Wnt (talk) 21:18, 15 January 2010 (UTC)

A link to the description page is the method of attribution that was chosen for the mediawiki software. The link can be removed if someone adds the name license next to the image in the text. Of course that would kind of defeat the point of using this image. What you are saying however, is that people who donated their works under a CC license shouldn't whine, because they know Wikipedia is good and that they should be lucky that Wikipedia will use their material. You say that Wikipedia has no moral obligation to think about how to respect the license in the best way we technically can at this point in time. I think a judge will say that you are passing by the rights of the authors a little bit too quickly here. I suggest we discuss the problem some more. —TheDJ (talk • contribs) 22:10, 15 January 2010 (UTC)

There is a link to the description page --- the "edit" link. When you press that link, a page comes up that directs you to it. The fact that it comes up as an "Image:Name" rather than directly loading is determined by how your the browser interprets the page - I think with Popups you could go there directly by clicking, actually.

Another way to link is

This provides an "info" button by default, though it seems to be in the way here. Wnt (talk) 05:31, 16 January 2010 (UTC)

What annoys me most about this suggestion that links should be banned is that CC-by-SA and GFDL multi-license is the recommended format for uploading to Wikipedia. In other words, a few months after the big vote to change the licensing of images on the site, we're stuck with some format that allegedly prohibits anyone from using cute little icons to link back and forth. Having previously encountered a longstanding bug that prevented template substitution in imagemaps, I get the feeling that Wikipedia is never, ever going to make links from images usable. Wnt (talk) 05:31, 16 January 2010 (UTC)

Wnt: If you are to lazy to make an image yourself (and release it as public domain so it doesn't need attribution), then you have to abide by the license and the law when you use images that others have made. And not attributing the author is rude. If you don't like the license, then don't use their image. It's as simple as that.

WFCforLife: Of course, this is a very simple image so it might not be copyrightable in the first place. If you don't know how to make images yourself and since it is such a simple image, ask the authors to change their license, or ask some image geeks to make a new one.

--David Göthberg (talk) 05:13, 16 January 2010 (UTC)

Heavens, let's not be lazy. Let's always remake every image every time we want to put up a page... For that matter, I suppose Wikipedia needs to stop serving pages like [26], which after all is what you get if you right-click and choose "copy image location"; this is the basis of the image links after all.

But seriously, you've just about convinced me: public domain is the only meaningful free license there is. We can fool around all day putting things up under weirdly named "free licenses" with cryptic names like CC-BY-SA 3.0 (who really pays attention to what that is anyway?), holding votes and debates and then find out years afterward that this doesn't mean simply that the image is attributed, but that it has to be displayed and linked in such-and-so a way. Meanwhile any spam site on the Internet, with more sensible sensibilities, can copy a whole Wikipedia page or part of it and put "from Wikipedia" in a corner and their obligations are done. It seems all the fancy legalese is just one big waste of time. Wnt (talk) 18:18, 16 January 2010 (UTC)

The fact that the 'bad guys' can do things easily and without effort, while the 'good guys' work harder to achieve things, is neither new nor surprising. People edit Wikipedia because they believe we are the 'good guys', and that what we are building is worth the extra effort. Spam sites may have more "sensible" morals, it's true. But then again, no spam sites are the sixth most-viewed site on the web. Happy‑melon 20:07, 16 January 2010 (UTC)

Yah, but the vote to change from the GFDL to the license that allows them to do that was made here! Wnt (talk) 02:10, 17 January 2010 (UTC)