Wikipedia:Village pump (technical)/Archive AU

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This page contains discussions that have been archived from Village pump (technical).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

Too much character lag

When someone vandalizes a page by adding a lot of characters it causes immense lag (1 minute to load a page). Could someone make a script so that when there is a specified amount of characters, the page doesn't show. -Flubeca^Talk 01:51, 21 August 2007 (UTC)

Ideally, the best way to implement this would be to examine the value of the Content-Length HTTP header, which MediaWiki uses. Unfortunately, it seems as though JavaScript doesn't provide a means of getting the Content-Length of the current page. XMLHttpRequest's getResponseHeader method can get the length of another page (with a HEAD request, probably), but that's not exactly helpful in this case. So that idea is perhaps out the window, and I can't think of something else that might work (other than using ajax to check a page before going to it). Perhaps there is a simpler solution :) Gracenotes^T § 04:13, 21 August 2007 (UTC)

Doesn't the javascript load after the page content? Prodego ^talk 04:33, 21 August 2007 (UTC)

I believe that depends on the script's position in the page. User scripts are loaded in the page's <head>, and may be executed before the DOM is constructed; onload hooks are loaded at the very bottom of the page (when the page is presumably loaded). Since HTTP/1.1, JavaScript can load concurrently with page content (and may even be cached). Gracenotes^T § 04:43, 21 August 2007 (UTC)

I just gave this a little test on test.wikipedia with XMLHttpRequest. The first problem I ran into is gzip. A 1 meg page of the number 9 (testwiki:1) is only 11469 bytes (according to Content-Length:) when gzipped, and most vandalism is probably very gzip-able. The next problem is, I don't think you can read the response headers (readyState 2) without actually loading the whole page into memory, so the bandwidth lag would be the same. An alternate method would be to XMLHttpRequest the history page and regexilly grab the byte value of the top edit (\<li\>$cur$.*$[0-9\,]{1,12} bytes$\<\/li\> or something). Flubeca: is your lag from slow bandwidth or from your browser attempting to render the page? --Splarka (rant) 08:17, 21 August 2007 (UTC)

If it's from rendering, you may want to check "Don't show page content below diffs" under the Misc tab in your preferences, or append "&diffonly=1" to the URL. Of course, your browser will still have to render to diff itself. —Ilmari Karonen (talk) 21:46, 21 August 2007 (UTC)

WP Size

What is the size of Wikipedia, in Mo?--SidiLemine 12:27, 20 August 2007 (UTC)

What's a "Mo"? --PEJL 12:49, 20 August 2007 (UTC)

The question presumably means in MB (or MiB). I'm not sure that there's a simple answer, because it depends on whether you mean just the English version, or all languages; whether you include all revisions of each article, or just the current one; images and other media or just text (and if you include images then the commons would need to be included); just articles or user and talk pages too. Do you mean the size of the wiki markup text, or of the generated html? The answers are to be found at Wikipedia:Database download and pages linked from there, although a simple answer for the simplest question of "html output of current english wikipedia articles only" is complicated by the dump being highly compressed. The current dump is running as I type this, so I can't see the file sizes, but it will be several GiB, highly compressed.-gadfium 19:37, 20 August 2007 (UTC)

Yeah, "Mo" = "Mega-octet", the common French term for Megabyte. Lupo 09:05, 21 August 2007 (UTC)

Job queue

Why is the current job queue length so long (about 2.7 million)? Aside from yesterday (when it was at about 1.5 million), I had never seen it go above 500 thousand and usually it's below 100 thousand. Of course, I don't check that often, but still ... — Black Falcon ^(Talk) 20:45, 18 August 2007 (UTC)

It fluctuates. Likely culprit is that someone made a series of changes to some widely used templates or categories. You also have to keep in mind that the job queue is not just for this wiki but for several (I don't believe all) Wikimedia wikis. In general, you shouldn't be considered by the length of the job queue unless you notice a lag occurring somewhere (populating a category, etc.). AmiDaniel (talk) 00:05, 19 August 2007 (UTC)

The job queue count shown on this wiki a very bad estimate of the length of this wiki's job queue only. Each wiki has an independent job queue, it's the job queue processing script that is shared between groups of wikis. 86.148.49.28 18:31, 21 August 2007 (UTC)

A lag (32 hours so far) in emptying a template-populated category is why I actually checked the queue. Of about 100 pages, only about 40 have been removed so far. It's nothing urgent, but I just wanted to know if perhaps something big was going on that I was unaware of. :) Thanks for your response, Black Falcon ^(Talk) 00:13, 19 August 2007 (UTC)

Never mind. The category apparently cleared out in the 3.5 hours since I posted my initial inquiry. — Black Falcon ^(Talk) 00:17, 19 August 2007 (UTC)

JeLuF fixed it. (He's a root). Cbrown1023 talk 01:02, 19 August 2007 (UTC)

The Hebrew Wikis

Does anyone knows why the Hebrew Wikiprojects (such as Wikipedia) fell down and currently cannot be edited? MathKnight 18:05, 14 August 2007 (UTC)

Database maintenance, most likely. Meta has been locked down for a good long while for maintenance as well (which is frustrating, since I've got something semi-important to post to a deletion discussion). EVula // talk // ☯ // 18:07, 14 August 2007 (UTC)

(followup) Yup, the message that is given translates to "Emergency maintenance, should be back in a few minutes", which is exactly what Meta says. Obviously not very accurate about the "in a few minutes" bit, but other than that, there's your answer. :) EVula // talk // ☯ // 18:10, 14 August 2007 (UTC)

A database server (the master for the cluster which the affected wikis use) crashed at about 17:20 (UTC). Standard procedure when things like this happen is to put the affected wikis into read-only mode to prevent any (further) data corruption or errors arising from MediaWiki attempting to write to a broken master. Brion Vibber recovered lost transactions and was able to bring the wikis back at around 18:18 (UTC). 86.148.49.28 18:54, 21 August 2007 (UTC)

Icons interfering with donations message...

Has anybody noticed that while you are logged out the icons indicating the status of an article like the semi-protection lock or the FA star overlaps with the "Donate to Wikipedia" message? It is happening in both IE and Firefox when logged out. The overlap can cause confusion and sometimes block out the donations text. --_Hdt83 _Chat 21:04, 20 August 2007 (UTC)

Extension:CategoryTree

Would it be possible to have $wgCategoryTreeMaxChildren set to 400? Cheers. --MZMcBride 19:30, 18 August 2007 (UTC)

Why? Is there a problem with its current setting of 200? Just because it uses AJAX and so you don't experience much lag in loading it doesn't mean that there's no burden on our servers, which is of course why it is set where it is. AmiDaniel (talk) 00:03, 19 August 2007 (UTC)

200 is an awfully low number, especially when you compare to the ability to see up to 5,000 transclusions in Special:Whatlinkshere or 5,000 contributions of a specific user. (Which isn't to say that they don't cause similar server strain; frankly, I have no idea.) 200 is the same as a standard view of a category page, but with the extension, there's no next button. Personally, it would be helpful with WP:DOT, which lists articles in a particular category, but has reached its limit. Cheers. --MZMcBride 00:42, 19 August 2007 (UTC)

You can just click on the category name, then you get the next button in the full-page view. -- 79.73.78.23 22:35, 20 August 2007 (UTC)

How can I disable the pop-up boxes?

Is this a new feature? Is there a way to turn it off? It's really anoying for me, because it interfers with my links. Please help me get rid of the pop-up boxes. It's good for newcomers I guess, and I'm sure others may appreciate it, but I think it should be an option in preferences. Can someone help me? - Jeeny ^Talk 04:57, 28 August 2007 (UTC)

You added it yourself by editing your monobook.js file. To remove it you'll need to remove the top four lines of text from that page. --Cherry blossom tree 08:17, 28 August 2007 (UTC)

Thank you! I'll do that now. :) - Jeeny ^Talk 08:21, 28 August 2007 (UTC)

Well, I deleted my monobook, but still have the issue. It must be a new update from MS. I do not have the problem while using FireFox. I don't know how to disable it in my IE preferences. Oh well. - Jeeny ^Talk 08:50, 28 August 2007 (UTC)

Clear your browser cache. —Ilmari Karonen (talk) 09:58, 28 August 2007 (UTC)

Image sizes as percentages?

Is it possible to introduce a feature that would allow images to be displayed as a percentage of the object containing them, as tables and their cells are? (Or is this a perennial proposal ....). Thanks ck lostsword•T•C 22:43, 26 August 2007 (UTC)

It's pretty tricky to do that; images are finicky. :) --brion 16:52, 27 August 2007 (UTC)

Ah well, thanks anyway Brion. I guess it'd be little use outside the User namespace anyway, so probably little point. Thanks once again :) ck lostsword•T•C 17:44, 27 August 2007 (UTC)

If you just want to do it in the user namespace, place it within a div and control it with CSS. Adrian M. H. 20:21, 27 August 2007 (UTC)

Please do this (new software changes)

MediaWiki:Cascadeprotected needs a $1 in there somewhere - to hold the list of pages. Needed because of my software changes in r24562. Thanks,

Werdna (editing from 58.172.200.115), 10:43, 26 August 2007 (UTC)

It seems to be working fine with $2, and I notice that Daniel changed $1 to $2 to apparently fix a problem. What exactly do $1 and $2 each do? Nihiltres^(t.l) 20:40, 26 August 2007 (UTC)

Whatever else is going on, apparently that template needs improved documentation. (SEWilco 04:24, 27 August 2007 (UTC))

The entire MediaWiki namespace needs improved documentation :) Gracenotes^T § 04:38, 27 August 2007 (UTC)

Well, hop to it :P AmiDaniel (talk) 17:27, 27 August 2007 (UTC)

Longstanding image/text space problem

There has been a serious issue in how text is formatted for some time. I assumed it would be fixed eventually but nothing has happened and it seems to be getting worse.

The problem is that if an in image is placed above some text and it runs into an image below it that was not placed directly below the first image tag in the markup, a void will be left in the text. In other words, if there is an image tag, then intervening text that will not necessarily fill the space until another image tag placed below the text, the text after the next image tag will appear in line with that next image. This behavior is wrong to me. Images should be formatted separately from the text and if there is not enough room to properly place an image it should just be stacked below the images above it.

This depends somewhat on display aspect ratio and resolution so it often goes unnoticed by editors with for example 1280x1024 displays. However on high resolution and widescreen displays it can become a severe formatting problem which leaves the article looking disjointed and very messy.

The problem behavior happens like this:


== First section ==

[[image:soandso.jpg|right|thumb]]
Some text that may not format out to be as tall as that image.

== Second section ==

[[image:another.jpg|right|thumb]]
This text will appear after an empty space (with any section heading left way above), at the position the another.jpg ends up.

Whereas this works properly:


== First section ==

[[image:soandso.jpg|right|thumb]]
[[image:another.jpg|right|thumb]]
Text associated with soandso.jpg

I have noticed this problem myself.. Does anyone know of a workaround? or what the issue is? --Quasar 22:55, 28 August 2007 (UTC)

Hrm, the archive bot got confused by the headers inside the <pre><nowiki> above. As posted earlier,

I would consider the behavior you describe to be broken as well, but fortunately it is just an Internet Explorer bug and not an issue with the existing CSS. Your "proper behavior" is exactly what occurs in Firefox, and reportedly in Opera and Safari as well. As noted, a potential workaround is to use {{clear}} or {{-}} just before the second header, which will move the gap to before the header instead of between the header and the text. Anomie 16:14, 16 August 2007 (UTC)

Anomie 23:36, 28 August 2007 (UTC)

Table/list with uneven borders

I created Template:Collapsible list collapsed because Template:Collapsible list will only collapse if more than two lists are present. The problem that I am having is that the borders are not even. The left and bottom borders seem to be thicker. Can anyone give it a second look to see if they see what I am doing wrong? Thanks —MJCdetroit 03:17, 23 August 2007 (UTC)

Why isn't this image working?

Compare:

http://upload.wikimedia.org/wikipedia/commons/thumb/c/c4/Gnome_Partition_Editor_showing_160_GB_disk.png/119px-Gnome_Partition_Editor_showing_160_GB_disk.png http://upload.wikimedia.org/wikipedia/commons/thumb/c/c4/Gnome_Partition_Editor_showing_160_GB_disk.png/120px-Gnome_Partition_Editor_showing_160_GB_disk.png http://upload.wikimedia.org/wikipedia/commons/thumb/c/c4/Gnome_Partition_Editor_showing_160_GB_disk.png/121px-Gnome_Partition_Editor_showing_160_GB_disk.png

— Omegatron 23:20, 22 August 2007 (UTC)

Oh, ok. You have to purge the image's description page. — Omegatron 23:23, 22 August 2007 (UTC)

Is there an error in footnotes?

The table in Warwickshire#Economy uses footnotes, but the list does not go 1, 2, 3 and 4 as expected, but it goes 1, 3, 4 and 5. Is this a bug? Snowman 16:31, 22 August 2007 (UTC)

Same issue as in this old discussion. The easy fix is to use <div style="font-size: smaller"> instead of <small> around the list. Anomie 16:49, 22 August 2007 (UTC)

The table in Macchi C.205 had got a static footnotes, because anchored footnotes did not work. The table has used [2] three times. If anchored footnotes are used the numbers ascend 2, 3 and 4 instead of staying as 2, 2 and 2. Is this a bug? Snowman 17:10, 22 August 2007 (UTC)

"This page is either protected or semi-protected."

Why isn't it possible to actually have the server (instead of the user) find out which kind of protection is applied (full or semi; permanent or temporary) and say so?

Right now the message is about as informative as "you can't edit it because you can't edit it". 24.83.195.130 13:35, 22 August 2007 (UTC)

Images in Faeroes

Why do several of the images in Faeroes not act as a link to the upload page? I thought all images did this. -- SGBailey 20:50, 21 August 2007 (UTC)

They are all fine for me. Adrian M. H. 21:11, 21 August 2007 (UTC)

(edit conflict) They seem to be fine on my screen. Both the images hosted on Commons and the images hosted on this Wikipedia. Valentinian ^{T / C} 21:12, 21 August 2007 (UTC)

I tried it again (using classic skin, XP & MSIE7) and it isa still the same. Flag OK, Coat OK, Map OK, Stamp OK, Tinganes BAD, Map OK, Aerial BAD, Litla BAD, Satellite BAD, Sheep OK, Ferry BAD, Church OK, Food OK, Flowers OK, Puffin OK. -- SGBailey 05:45, 22 August 2007 (UTC)

I think you need to purge your IE cache. I just rechecked and the images are fine on my screen, including the image description pages from Commons. Valentinian ^{T / C} 05:49, 22 August 2007 (UTC)

Interesting. I purged my cache and tried it in Monobook - no problems. I logged in and used Classic skin and the problems recur. Log out and ok in monobook. -- SGBailey 07:05, 22 August 2007 (UTC)

Exactly the same thing happens withthe two skins on a different PC (at work) -- SGBailey 08:07, 22 August 2007 (UTC)

template show / hide (showing by default when should be hide)

I just noticed that many of the templates that I watch that have a show / hide function (that hide by default) are now all showing by default. For some reason, the show is expanded when it should be hidden on page load. Examples {{public finance}}, {{UStaxation}}. Morphh ^(talk) 13:12, 08 August 2007 (UTC)

This is a result to the NavFrame bug fix applied to Common.js. They used to collapse if there were three or more on a page, but now they are always open unless the NavContent style="display:none;..." (so you now have control over NavFrame's initial state.) Many of the NavFrames can be traced back to a protected template, many of which have {{editprotected}}s outstanding. I will fix all that I can see. If anyone sees more, list them here, or if you feel ambitious, add the display:none; to the NavContent style. ←Ben^B4 13:47, 8 August 2007 (UTC)

Fixed the two above. Morphh ^(talk) 13:58, 08 August 2007 (UTC)

Thanks, you beat me to it! ←Ben^B4 14:20, 8 August 2007 (UTC)

{{WikiProject Taxation}} may also have this issue. Morphh ^(talk) 13:53, 08 August 2007 (UTC)

~~Can you give an example page, please? I can't find any NavFrame uses.~~

Done ←Ben^B4 15:48, 8 August 2007 (UTC)

I noticed this issue on a couple templates (including some I created) as well. The thing about them is they are set to show if a certain parameter is set. Please take a look at {{Johnny Cash}}, {{University of Oklahoma}}, and {{University of Texas at Austin}}. For instance, for the last two, they have around 4 or 5 separate dropdowns. But if the "athletic" parameter was set, then the athletics drop down would be displayed where all others would be hidden. Don't know if this functionality has been affected by this change.↔NMajdan•talk 13:20, 13 August 2007 (UTC)

Tried to fix it (so now there all hidden), though I don't think it fixed the option that would show the one selected. This is a little beyond my skill. Morphh ^(talk) 18:56, 14 August 2007 (UTC)

Ok, finally had some time to look myself. I think I got it fixed. I have fixed the OU template and will tackle to other two next week unless somebody beats me to it. Here are the changes that resolved the issue on the OU template.↔NMajdan•talk 19:19, 16 August 2007 (UTC)

Fixed the UT and Johnny Cash templates today.↔NMajdan•talk 17:31, 22 August 2007 (UTC)

As of now, the section that is supposed to show does not show at all and the [Show] button is disabled :o Corpx 00:22, 15 August 2007 (UTC)

Those all work for me. ←Ben^B4 00:57, 18 August 2007 (UTC)

I think that the {{Royal Navy ships}} template seen at Royal navy has the same problem. Woodym555 17:45, 16 August 2007 (UTC)

Fixed. ←Ben^B4 00:48, 18 August 2007 (UTC)

Excellent, thanks, it was taking up half the page! Woodym555 10:25, 18 August 2007 (UTC)

Monobook not working

Wikipedia is no longer displaying in monobook for me, logged in or not. Instead I seem to be stuck in "myskin." Changing my preferences allows me to change to other skins like classic and blueskin, but when I toggle back to monobook, WP still displays in myskin. Is this happening to anyone else? Thanks, Dar-Ape 22:25, 29 August 2007 (UTC)

Nevermind, the problem seems to have gone away... Dar-Ape 18:17, 30 August 2007 (UTC)

Section headings and infoboxes

I just posted this topic at the help desk and was told to post here. This might be a bit of an obvious issue, but is there any way the CSS of Wikipedia can be changed so that the dividing lines underneath headings don’t clash with infoboxes or right-aligned templates? It creates a real visual clash on lots of pages, surely their must be a workable solution. Sorry if this question has been asked before. Max Naylor 17:07, 27 August 2007 (UTC)

Clash in what way? Do mean that there is some overflow? Normally, the lines should stop at the box or image. If you have problems with misplaced edit links, see WP:BUNCH. Adrian M. H. 17:54, 27 August 2007 (UTC)

One solution would be to use the template {{-}} to clear elements from "floating" next to each other. This may create large amounts of blank space, however. Gracenotes^T § 19:12, 27 August 2007 (UTC)

No, I‘m not looking to solve the bunched up links problem, for example in Culture of Iceland the underscore of the section headings goes straight through the image, although behind it (you might have to have a wide screen resolution to make it work). I have posted a screenshot here [1]. I hope this makes things clearer. I don’t think it’s a browser problem, however if it helps I am using Safari on Mac OS X (I have looked at the same page in Firefox and the problem still persists). Max Naylor 12:50, 29 August 2007 (UTC)

Take a look at nl:Gebruiker:Erik Baas/Test; it demonstrates that section headings like "=Heading 1=" and

Template:E

Template:E has a line break in it that messes articles up. Is there a way of getting rid of this line break, or is there a reason for it being there - Inwhich case can we make a Template:EWithNoLineBreak? This question was firstraised on its talk page last year and no reply or fix has been done. -- SGBailey 08:29, 28 August 2007 (UTC)

Fixed. —Ilmari Karonen (talk) 10:32, 28 August 2007 (UTC)

Thanks -- SGBailey 08:37, 29 August 2007 (UTC)

Help needed with complex template

I've created Template:Ancient monuments in Rome (based on a similar French and Dutch template) to serve as an infobox. However, I'd like to extend it using some of the labeled map template syntax. As an experiment, I created Template:Ancient Rome Labelled Map Cestius. You can see how this works - it overlays a marker and text on top of an image (in this case a map) at a set of x,y coordinates. I then transcluded this into a cell within Template:Ancient monuments in Rome. You can see the two templates in use at Pyramid of Cestius. I'd like to incorporate the x,y coordinate fields into Template:Ancient monuments in Rome so that any point in the city can be marked, without having to create a separate template or map for each location. However, I don't know enough about template syntax to be able to do this. Can anyone help? -- ChrisO 20:23, 24 August 2007 (UTC)

I've tried to generalize the map function, and it seems to work. (The {{image label}} system has a couple of unintuitive idiosyncrasies, unfortunately.) Gracenotes^T § 22:10, 24 August 2007 (UTC)

Thanks very much! It needed a couple more tweaks, but it's working fine now. -- ChrisO 00:14, 25 August 2007 (UTC)

gets the users’ list of a particular DataBase when name of the DB is supplied through query

I need to know the SQL query that gets the users’ list of a particular DataBase when name of the DB is supplied through query. Please treat this as important and do pass the information as soon as possible —The preceding unsigned comment was added by 203.99.204.99 (talk) 09:40, August 21, 2007 (UTC)

What?? —Preceding unsigned comment added by Orngjce223 (talk • contribs) 03:50, August 25, 2007 (UTC)

This is the wrong place for this kind of question. Try the reference desk. Just a hint: you'll need to be more specific than this. Different RDBMSes handle this kind of thing differently. TCC (talk) (contribs) 04:00, 25 August 2007 (UTC)

Schadenfreude article first line displays funny

I appreciate what the previous editors were trying to do by incorporating/linking the audio version of the word Schadenfreude in the first line of article, but the display looks funny on my machine (others too?). Also, I am not sure that this format agrees with WP:MOS. Comments?--Filll 18:28, 24 August 2007 (UTC)

Watchlist stuck at 14:51 GMT?

Has anybody else noticed their watchlist stuck at 14:51 GMT? No amount of purging or changing settings helps. Agathoclea 15:39, 24 August 2007 (UTC)

It's happening to me, too. Someone go wake up the hamsters... --ElKevbo 15:41, 24 August 2007 (UTC)

Me, three. I imagine that it's related to the bad software update that occurred around then. I guess watchlists are low priority for the system resources as they get everything else back. Flyguy649 ^talk _contribs 16:03, 24 August 2007 (UTC)

Me four. I'm surprised it's still stuck more than an hour later, though. —Angr/talk 16:08, 24 August 2007 (UTC)

Should be fine now. We had some fun with databases when upgrading the software today. --brion 17:58, 24 August 2007 (UTC)

Watchlist notification thingy

I noticed (yesterday or the day before) one of those notification thingies at the top of my watchlist. I didn't dismiss it, yet it's disappeared. What was it for? How comes it disappeared so quickly? --Dweller 10:51, 24 August 2007 (UTC)

Probably it was dismissed as a bad idea. It was a similar situation with the Add a message tab that replaced the + tab at the top of talkpages for about half-a-day - that was something which I actually quite liked! That's all it'll be I think. Lra drama 10:56, 24 August 2007 (UTC)

It was inviting me to participate in some debate or vote and I never got round to it. It wasn't a new bit of kit, just one of the regular notification type messages we get from time to time, like when there are elections. --Dweller 11:02, 24 August 2007 (UTC)

It is quite normal for notices of that nature to be removed quickly because some editors use the watchlist notice for things that do not really warrant it. WP has other paths for highlighting discussions. Adrian M. H. 11:08, 24 August 2007 (UTC)

OK... anyone know how I can find out what it was? --Dweller 11:38, 24 August 2007 (UTC)

Any sitenoice/watchlist notice changes correspond to a MediaWiki page notice. I always go to recenchanges and filter by MediaWiki namespace to track such changes. Voice-of-All 12:41, 24 August 2007 (UTC)

It was about Flagged revisions if I recall correctly. Adrian M. H. 12:45, 24 August 2007 (UTC)

Oh. Is it just the jargon, or is it poor English as well, that makes my brain spin when I read the first parag of that article? --Dweller 13:06, 24 August 2007 (UTC)

The message that you are talking about can be found here. Watchlist messages are modified at MediaWiki:Watchdetails. I found this out by doing a google search for "watchlist intitle:MediaWiki". Graham87 14:02, 24 August 2007 (UTC)

Nice one. --Dweller 14:04, 24 August 2007 (UTC)

5.1 Speakers + Xbox 360 + HD/VGA Cable + PC Monitor??!?!?!?!?!?!

I have got a Creative SBS 580 5.1 surround sound system and i want to connect my xbox 360 to the speakers and my pc monitor! Therefore i bought the HD to VGA cable, the thing is, my speakers have three cable coming from them, an orange, green and a black one while the cable has two things for audio, the first is a red and a white cable and the second is a digital audio input! So what will i need to connect them all? —Preceding unsigned comment added by 172.188.189.83 (talk) 08:46, August 24, 2007 (UTC)

Please read the header at the top of this page, then phone Microsoft. Adrian M. H. 10:54, 24 August 2007 (UTC)

New donation links

The new donation links seem to appear twice in Firefox 2 (in main namespace), once above the tabs (with "continued donations" not in bold), one (with "continued donations" in bold) where it normally is. Why? 203.214.76.82 07:30, 24 August 2007 (UTC)

Watchlist - Individual section watching & "Special Aritcles"

On many pages (mostly like this one, or noticeboards, or WP:PG discussion pages) - is there anyway to actually monitor specific sections of the content? Many of the times, when people make updates to a discussion, it will get "lost in the shuffle" on my watchlist. Additionally, there are certain articles that i would love to be able to "monitor" seperatly. Whether this was a "subsection" of my watchlist, or a special page that is created - is there someway to have "articles of special attention"? Juan Miguel Fangio| ►Chat 04:53, 24 August 2007 (UTC)

Per your second question, you can create a page in your user space, for example User:Jmfangio/watchlist_2, and link to any articles on it you want to give special attention to (but not put on your watchlist). On that page will be a link (as is on every page) to Special:Recentchangeslinked, which is a cropped version of Special:Recentchanges, only showing changes in the pages linked to on the page in question. For example, the related changes for this pump are at Special:Recentchangeslinked/Wikipedia:Village_pump_(technical). --Splarka (rant) 07:32, 24 August 2007 (UTC)

Interesting work around. Never thought of that. I actually would use that for additional "watching" of certain articles. Since it appears that this functionality is "hidden" in the system - how difficult would it be to get something like this worked into the functionality of the wantlist (like a filter of sorts)? Juan Miguel Fangio| ►Chat 07:59, 24 August 2007 (UTC)

Template:08 ?

I was fooling around with this userbox: {{User current age}} and I happened to notice a redlinked template called {{08}} on the edit preview pages where it is used. Can someone please explain to me why this empty template is being transcluded to that user box? It is transcluded on just under 1000 pages and I can't figure out where/how it is used so I can get rid of it. Am I missing something? Thanks very much. Paul^C/_T+ 21:06, 23 August 2007 (UTC)

Oh, I forgot to mention, there has been some discussion about this on the talk page of the template already... but noone knows what is going on. Paul^C/_T+ 21:09, 23 August 2007 (UTC)

Commented on the template talk. --TMF ^{Let's Go Mets - Stats} 21:24, 23 August 2007 (UTC)

Replied on talk. It looks like {{Days in month}} had been broken since May 5, 2006. Anomie 01:19, 24 August 2007 (UTC)

Is there a way to change `wgContentLanguage`?

I have a lot of questions around this, but for now I'll go with this one: I saw that following a link which has ?uselang=xx changes the value of wgUserLanguage (can be seen on the source code of the page), which changes the language of the tabs and other things. For example, if you follow this link you'll see that the tabs are in basque, and if you look at the source code, the variable wgUserLanguage will say eu. Now, is there anything to change the value of wgContentLanguage? I want this because it's the only one which has a magic word to get its value, {{CONTENTLANGUAGE}} precisely, and there's no magic word to get wgUserLanguage, which btw I don't know why. Thanks for your help. - Keta 09:44, 23 August 2007 (UTC)

This is presumably by design: you'll notice that there are no other magic words that provide user-specific information either. This is because pages may be parsed in advance and cached internally, so the parser doesn't really have reliable access to that information. —Ilmari Karonen (talk) 16:22, 23 August 2007 (UTC)

Hmm, I think I don't get you. Using the link with "?uselang=xx" changes my user language even if I have another one preset. You mean this magic word wouldn't get the new value? - Keta 22:17, 23 August 2007 (UTC)

There is a similar thread above outlining why such magic words would be troublesome to implement on a Wikimedia wiki. You'll probably have to use a javascript-based solution (as above). It would be handy to be able to be able to

{{#ifexist:{{FULLPAGENAME}}/{{USERLANGUAGE}}|{{{{FULLPAGENAME}}/{{USERLANGUAGE}}}}|{{pleasetranslate}}{{{{FULLPAGENAME}}/en}}}}

which could then pull a translation of the current page from the appropriate subpage, or ask the user to translate it by creating a subpage in their language. But currently this would break caching. --Splarka (rant) 23:45, 23 August 2007 (UTC)

Thanks Splarka. Is there any chance that in the future something like this could be used? Not exactly this, but something which would not need javascript on each wiki? - Keta 10:48, 24 August 2007 (UTC)

If I use the "back" button on my browser, all Wikipedia's button's stop working!

If I'm viewing history on a page, I used to be able to hit the "compare" button, view the changes, hit the "back" button on my browser, choose a different set of revisions, hit "compare" again, go back, until I was satisfied I understood what had been done to the article (or had found the revision that inserted a particular comment I was curious about). But for a week or so now, this no longer works - to get the "compare" button to work, I have to reload the page.

Same thing for the buttons at the bottom of the editing screen. If I'm unhappy with how an edit looks in preview mode, and want to go back to a previous edit (that had not been saved to Wikipedia yet, but was still in my browser's history) - I can do that, but have to copy and paste text into a reloaded window to be able to preview or save that information!

Any idea what happened? Lyrl^Talk _C 23:41, 21 August 2007 (UTC)

Browser? Any recent updates to your system? Fiddle with any preferences settings? TCC (talk) (contribs) 11:16, 24 August 2007 (UTC)

Non-free software screenshot categorization

Comments at MediaWiki talk:Licenses/en-nonfree would be appreciated. —Remember the dot ^(talk) 00:47, 26 August 2007 (UTC)

Twinkle problem

I think there's been some technical update to the deletion system for files, and it's messed up Twinkle for one, which makes not as able clear out backlogs (especially the ones that require removing images from articles, and the fact I don't have to type). I've been searching for it ever since it basically interrupted me mid deletion (one part of my tabbed images were deleted, the other ones not.) I've gone as far as deleting and undeleting the MediaWiki messages for deleting the file and restoring it. I have run out of ideas. Please help! Maxim_(talk) 22:21, 25 August 2007 (UTC)

There was just a massive software update, focusing on the image namespace (see a couple of posts up). This is most likely a symptom of that. I suggest leaving a message at Twinkle talk. Cheers. --MZMcBride 22:29, 25 August 2007 (UTC)

I've tried to contact the author (AzaToth), but I don't I've had a reply.

We've recently fixed some problems with file undeletion, but haven't run across deletion problems yet. We'll keep an eye out; it might help if you can point out examples of files which are problematic. --brion 22:30, 25 August 2007 (UTC)

Any image deletion won't work with twinkle. That what's problematic. Maxim_(talk) 22:42, 25 August 2007 (UTC)

The user interface has changed, so third-party scripts which screen-scrape the UI probably have to be updated. --brion 22:54, 25 August 2007 (UTC)

Guess someone else will have to clear image deletion backlogs for the next while. Oh well. Maxim_(talk) 22:58, 25 August 2007 (UTC)

Does this page look weird, or is it just me?

http://wikipediaweekly.org/category/audio/mp3/feed

--Filll 16:24, 25 August 2007 (UTC)

Guess it's just you; looks like a normal enough podcast RSS feed to me. --brion 17:42, 25 August 2007 (UTC)

MediaWiki:Accesskey-pt-preferences

I was looking through the recent software changes for the Signpost, and discovered that the default value for this message (the access key for Special:Preferences) was being changed from 'nothing' to ','. There's no problem with this; the problem is that the message currently seems to have the value '9', for no reason at all; it has no edit history and no deleted versions, so I can't think of anywhere its current value could have come from (it's not in the relevant file in the software that determines the default values of MediaWiki messages). Does anyone know how the message managed to get that value? Will the new software change overwrite the current value of '9', or just not take effect? --ais523 16:20, 25 August 2007 (U T C)

According to the commit logs it was changed to 9 because , conflicted with something. --brion 17:41, 25 August 2007 (UTC)

What happened to image sizing?

Resolved

It used to be that [[Image:I-95.svg|25x20px]] returned a 20px by 20px image, since that was the largest that fit into the specified rectangle. But now it gives 25px by 25px: Is this related to the above update? If so, and it's not a bug, what's the new format to specify height? --NE2 14:34, 25 August 2007 (UTC)

Sounds like a bug, I'll take a peek. --brion 15:27, 25 August 2007 (UTC)

Thank you very much, and good luck. --NE2 15:51, 25 August 2007 (UTC)

Fixed. --brion 15:53, 25 August 2007 (UTC)

If any are still bad, purge. --NE2 15:57, 25 August 2007 (UTC)

Software update

Ok, we finally got the software updated again, this time without breaking the database servers. ;)

Various bugs should be fixed, and file info display & reversion should be a little nicer. Let us know if anything's broken worse than before. --brion 14:12, 25 August 2007 (UTC)

The "Upload file" link still does not work when using the secure proxy. —Remember the dot ^(talk) 15:52, 25 August 2007 (UTC)

Link names on page history/user contribs

(copied from Village pump (proposals))

Compare these:

User contribs: (Newest | Oldest) View (Newer 50) (Older 50) (20 | 50 | 100 | 250 | 500).
Page history: (Latest | Earliest) View (previous 50) (next 50) (20 | 50 | 100 | 250 | 500).

It would probably be better if both used the same names, preferebly not "previous" and "next" which mean exactly the opposite of what one would expect in the page history. Any thoughts? Zocky | picture popups 09:13, 24 August 2007 (UTC)

It does not matter in the slightest since the meaning is still clear. Adrian M. H. 09:18, 24 August 2007 (UTC)

I don't think "previous" (which means "later") and "next" (which means "earlier") are very clear. -- Jitse Niesen (talk) 11:00, 24 August 2007 (UTC)

I always hit the wrong one when it's "previous" and "next". Thanks for drawing attention to this problem, it never occurred to me that it needed fixing.--Father Goose 15:28, 24 August 2007 (UTC)

I've been confused by this on numerous occaisions, I agree that it should be standardised, and I agree with BigNate below. SamBC(talk) 17:44, 24 August 2007 (UTC)

Standardizing on new(er|est)/old(er|est) would seem logical. Can we have this changed via a MediaWiki page, or does this need to be changed behind the scenes? BigNate37_(T) 16:58, 24 August 2007 (UTC)

In newer versions of the MediaWiki software it's easy to customize these kinds of messages (you enter something on a page overrides the default messages). The reference is somewhere around Meta:Help:System message. —Preceding unsigned comment added by Wikidemo (talk • contribs) 19:41, August 24, 2007 (UTC)

I'm going to transplant this thread to Wikipedia:Village pump (technical) to get the developers' attention, since it should be fixed in the distribution, not just by user customization.--Father Goose 20:19, 24 August 2007 (UTC)

Submitted to bugzilla.--Father Goose 23:07, 25 August 2007 (UTC)

overlap weirdness??

(overlap weirdness??)

Lately I've been getting odd overlaps with the banner at the top of the page when looking at minimized windows, can someone else confirm that it's not just me?--69.118.235.97 18:33, 30 August 2007 (UTC)

Yeah, we were trying out a new way of placing the donation notice above the left tabs, as opposed to in the upper-right, and it's definitely having issues on several different browsers with, exactly as you describe, text overlap. I've now stuck her back in the upper-right, and hopefully we can get the kinks ironed out. AmiDaniel (talk) 21:59, 30 August 2007 (UTC)

(Still) getting some odd overlap issues.--69.118.235.97 22:44, 30 August 2007 (UTC)

Please stop cropping your screenshots so close in. It makes it hard to get an idea of why everything is wrapping like that. So just take a screenshot of the entire window, and that will help us to help you. --Cyde Weys 19:40, 2 September 2007 (UTC)

Metadata problem solved

A couple of weeks ago a user reported that the image metadata was linking to the wrong place, i.e., the automatic linking of the camera manufacturer was causing issues. I've solved the problem by modifying MediaWiki:Exif-make-value and creating Template:Exif-make-value. Any automatic metadata can now be #switch'ed to link to the correct place. So, instead of having a redirect bounce from the default "OLYMPUS IMAGING CORP." to "Olympus Corporation", it's now is linked correctly in the first place. Example image. Cheers. --MZMcBride 04:18, 27 August 2007 (UTC)

"This is a User Page" technical/obfuscation question

Hello, VP followers. I have asked this question of a couple other Wiki users, and have found it easier to create a page that asks my question instead. Can anyone answer for me, either here or on my page? TIA. --Mike Vitale 21:08, 26 August 2007 (UTC)

It's done to prevent mirrors from automatically replacing the word "Wikipedia" with the mirror site's name. Mike Dillon 21:20, 26 August 2007 (UTC)

(edit conflict) It has to do with Wikipedia mirror sites, pretty much. This template history will give you a pretty good idea (specifically these diffs: [2] [3]. See also: Template talk:Userpage#Escape sequencing. Cheers. --MZMcBride 21:24, 26 August 2007 (UTC)

Thanks to both of you. I assumed it was something along those lines. Bleah, I hate obfu; reminds me of my days writing Perl. --Mike Vitale 00:45, 27 August 2007 (UTC)

"What links here"

Hey guys - Sorry to bother you but I wish to know what order the articles are in the "What links here" option.--Phill ^{talk Edits} 12:22, 3 May 2007 (UTC)

They aren't ordered. There must be something that determines the order, but it's based on number of links, time of addition, phase of the moon... So it's effectively completely random. -Amarkov moo! 13:46, 3 May 2007 (UTC)

They are ordered by the time of the article's creation - oldest first - as stated rather cryptically at help:what links here#Order; query. So that means, for example, that sun would probably be listed as one of the first while Wikipedia:Today's featured article/May 4, 2007 would be listed near the end of the list. You can see this in action at special:whatlinkshere/Bureau of Meteorology; I recently added a link to the Bureau of Meteorology article to Darling Scarp but Darling Scarp appears near the start of the list because it was created in November 2002. Graham87 13:54, 3 May 2007 (UTC)

This would have been nearly impossible to find in the Village pump archives unless you knew what to look for. Cheers. --MZMcBride 21:16, 26 August 2007 (UTC)

Brilliant, thanks! Unfortunately, that makes my task much more difficult! Thanks for your speedy assistance though. Verisimilus T 00:06, 27 August 2007 (UTC)

What links here?

Hi,

I'm wondering whether new uses of a template are always added at the end of a "what links here" list, for bot-related reasons.

As I struggle to find my way around the back waters of Wikipedia, is there somewhere I could have found this out myself?

Thanks,

Verisimilus T 21:03, 26 August 2007 (UTC)

From Wikipedia:Village pump (technical), May 3, 2007

Using the "orphanBacklinksOnSpeedyDelete" feature on Twinkle

I would like to remind admins who are using the Twinkle user script to be careful when removing backlinks to pages that they have speedy deleted using this script. Sometimes a vandal or new user will click on a legitimate red link and create a speedy deletable page. Please remember that WP:RED says that "removal of red links for nonexistent topics should not be done without careful consideration of their importance or relevance." Thanks. Zzyzx11 (Talk) 05:01, 26 August 2007 (UTC)

I've put in a feature request to make deleting backlinks an option in the CSD menu. Mr.Z-mantalk¢ 20:45, 26 August 2007 (UTC)

html to pdf

can anyone know an extension to export wiki pages into pdf format?? —Preceding Naveenpf 04:43, 26 August 2007 (UTC)

Why don't you just print to pdf? Juan Miguel Fangio| ►Chat 04:45, 26 August 2007 (UTC)

Probably because, like most people, he doesn't own a copy of whatever software you have installed that adds PDF generation as a printer. So it would help to mention a GPLed package that can do this, like PDF Creator. (Which I haven't tried yet myself, but having had to do a search by way of figuring out what you were probably talking about, looks pretty useful if it works as advertised. TCC (talk) (contribs) 20:06, 26 August 2007 (UTC)

I have not tried this, but [4] might be what you're looking for. Adrian M. H. 11:01, 26 August 2007 (UTC)

PDF output is discussed at Wikipedia:WikiReader#Resources; it's not clear if any of the options listed are good/easy solutions. -- John Broughton (♫♫) 23:31, 26 August 2007 (UTC)

Notification of page moves

I find it somewhat unusual that when a page i watch is moved, it does not reflect this on my watchlist. How difficult would it be to build out some sort of notification tool for this and would that be appropriate? Juan Miguel Fangio| ►Chat 07:57, 21 August 2007 (UTC)

This is a rather old bug, due to certain page actions being given a revision (so as to appear in the page's history and not just the logs for the page), which then doesn't jive with Special:Watchlist 100%. Or something. That is a confusing bug debate. --Splarka (rant) 08:24, 21 August 2007 (UTC)

D'OH!!! :-) Juan Miguel Fangio| ►Chat 15:07, 21 August 2007 (UTC)

There is a workaround - you can add User:Dschwen/HighlightRedirects and then use it to turn all redirects on your watchlist green. (I'm assuming - perhaps incorrectly - that your watchlist will (still) link to what is now a redirect page rather than link directly to the new location of the page.) -- John Broughton (♫♫) 23:43, 26 August 2007 (UTC)

cite.php

I spend a lot of time fixing citation problems. I wonder if it might be possible to enhance cite.php processing to detect two error conditions and, if those conditions are detected, to request confirmation before saving similarly to what is done if there is an attempt to save without an edit summary. The two conditions are:

an attempt to save a page containing one of more <ref> tags, but not containing a <references/> tag.
an attempt to save a page where there is a null expansion of a <ref> tag. This commonly occurs where article text containing a <ref name=whatever>...</ref> tag is deleted, leaving behind one or more orphaned <ref name=whatever/> tags. -- Boracay Bill 05:52, 20 August 2007 (UTC)

Unless I'm gravely mistaken, cite.php runs when a page is served (displayed), not when it saved. Also, how is this supposed to work with section editing? You could just use preview to catch both types of errors. Lupo 10:40, 20 August 2007 (UTC)

Pages are rendered to HTML on save, including <ref> tags. Sections are merged back into the rest of the article. It's a reasonable feature request. -- 79.73.120.112 11:54, 20 August 2007 (UTC)

That's news to me, and I don't believe it'd work. Think about template expansions, and changes to templates. You cannot "expand to HTML" upon save. If the database stored only expanded HTML, you'd lose any template references, and couldn't update them when a template changed. Hence, store the wiki text, and expand to HTML upon serving the page, then that referential problem disappears. See also [5]: the text table stores wikitext. The caches, now they do cache rendered pages. Which is why you sometimes have to purge a page when changed transcluded content is not displayed. Lupo 09:03, 21 August 2007 (UTC)

Lupo, the citation problems I speak of spending a lot of time fixing above are problems which I discover in pages edited by others who, for whatever reason, have not caught those problems prior to saving the pages they've edited — leaving those problems behind in the saved page. The objective of this feature would be to warn people who are about to inadvertently save pages with these sorts of detectable problems in them that they are about to save a page with these sorts of detectable problems in them. -- Boracay Bill 04:58, 21 August 2007 (UTC)

Like the wikitext is expanded in preview, it could be expanded when pressing "save" (before actually saving) for the purpose of this check.--Patrick 10:10, 21 August 2007 (UTC)

That is true. The software could parse the wikitext and note any problems it detects before saving, switching to preview mode if any problems were found. I wonder what the impact on the performance would be... simply storing pages is probably not the most crucial bottleneck, but to expand a page, you'd also have to read (potentially a lot) from the DB, and you couldn't use the caches. If feasible performance-wise, such a check before saving could also be used to catch other problems, such as typing "{{Category:XYZ}}", or "[internal link]]", or "[[internal link}}", or unclosed refs ("<ref>blah blah blah ... more text ... <ref>foo bar</ref>"), or "=== Some title ==", or ... The list of checks to do could maybe even be a per-user preference setting, with defaults for anonymous users. And some types of mistakes could maybe even be auto-corrected, such as changing "[[internal link}}" to "[[internal link]]", or even to replace "the the" by "the" and "who who" by "who"... Who's gonna code it? Lupo 10:31, 21 August 2007 (UTC)

File a feature request, asking for error checking be added to MediaWiki. Then learn how to program MediaWiki and write the code so the task will get done. (SEWilco 23:34, 21 August 2007 (UTC))

Typical smug "sofixit" answer. Doesn't work. I can program—it's what I do for a life—but I don't have the time to start MediaWiki development on the side. Lupo 06:20, 22 August 2007 (UTC)

(undent) An alternative would be to post this at Wikipedia:Bot requests. While getting a bot to check for this problem isn't anywhere near as satisfying as getting the developers to change the software, it's much more likely to happen.

And if you do post a request for a bot to be coded (or modified) to do this, you'll need to specify what you want the bot to do. In the first situation, presumably the bot would create a "References" section (do note that Template:reflist and, for all I know, other templates, provide the same functionality as <references />, so a bot needs to check for all of these). In the second situation, probably the best you're going to get is for the bot to post a note to the article's talk page. (Ideally, it would identify the edit that caused the problem, but that might be asking a lot of a bot.) -- John Broughton (♫♫) 23:57, 26 August 2007 (UTC)

Block log problems

Something seems to be wrong with the block log: this user is blocked, though their block log is empty. —Ilmari Karonen (talk) 02:48, 28 August 2007 (UTC)

Blocked text not working?

I was recently on a shared computer at the local library and (as expected) the computer was blocked from editing Wikipedia. What was interesting was that instead of giving out the usual MediaWiki:Blockedtext message, it was giving out a generic block message. Is something wrong with teh blocking system? Or is it just that computer? Can someone verify this... --_Hdt83 _Chat 22:44, 27 August 2007 (UTC)

Image sizes as percentages?

It's pretty tricky to do that; images are finicky. :) --brion 16:52, 27 August 2007 (UTC)

Ah well, thanks anyway Brion. I guess it'd be little use outside the User namespace anyway, so probably little point. Thanks once again :) ck lostsword•T•C 17:44, 27 August 2007 (UTC)

If you just want to do it in the user namespace, place it within a div and control it with CSS. Adrian M. H. 20:21, 27 August 2007 (UTC)

Please do this (new software changes)

MediaWiki:Cascadeprotected needs a $1 in there somewhere - to hold the list of pages. Needed because of my software changes in r24562. Thanks,

Werdna (editing from 58.172.200.115), 10:43, 26 August 2007 (UTC)

It seems to be working fine with $2, and I notice that Daniel changed $1 to $2 to apparently fix a problem. What exactly do $1 and $2 each do? Nihiltres^(t.l) 20:40, 26 August 2007 (UTC)

Whatever else is going on, apparently that template needs improved documentation. (SEWilco 04:24, 27 August 2007 (UTC))

The entire MediaWiki namespace needs improved documentation :) Gracenotes^T § 04:38, 27 August 2007 (UTC)

Well, hop to it :P AmiDaniel (talk) 17:27, 27 August 2007 (UTC)

Redirects with + don't work

A problem was reported at Wikipedia:Help desk#Redirect fails to move user to article. I have not found any working redirect to a target containing the + character. The reported example contained other special characters. A simpler example: Channel 1+1 contains #REDIRECT [[1+1]] so it should redirect to 1+1 but it doesn't. The link to 1+1 works in this post and it works in preview at Channel 1+1, but not in the saved redirect page where + in the link is replaced by a space. PrimeHunter 02:22, 26 August 2007 (UTC)

This is easily fixed by percent-encoding the + as %2B. I fixed the page in question, by changing the target of the redirect to "1%2B1". Nihiltres^(t.l) 02:30, 26 August 2007 (UTC)

I would say that this is a bug. It should be reported at bugzilla. Mike Dillon 02:55, 26 August 2007 (UTC)

I also considered that. There might be thousands of broken redirects with +. I don't know a way to search for them and new ones can be created. I guess the problem was not there when most of the existing redirects were made. I don't have experience with bugzilla. Will somebody search to see if it's already there and report it otherwise? PrimeHunter 02:59, 26 August 2007 (UTC)

It would be possible to go around with a bot, fixing every redirect with a + in it. But that's not that most practical solution :) Gracenotes^T § 03:29, 26 August 2007 (UTC)

It's probably a side-effect of bug 10683 (that is, the fix applied to the bug, which became active very recently.) Nihiltres^(t.l) 04:25, 26 August 2007 (UTC)

Definitely. The urldecode is turning "+" into a space, as it should. I don't think that bugzilla:10683 should have been fixed; I don't see why URL encoding should be allowed in wikilinks, since they aren't URLs. Mike Dillon 14:48, 26 August 2007 (UTC)

To match existing inline linking behavior. --brion 14:56, 26 August 2007 (UTC)

I meant in inline links too :) Mike Dillon 21:21, 26 August 2007 (UTC)

We try not to gratuitously break existing markup features, especially on a site with millions of pages. ;) It'll remain for backwards compatibility.

The primary usage requiring it in the olden days was interlanguage links; before we transitioned everything to UTF-8 a couple years back, putting in the percent-encoded strings was the only way to specify a link to a non-latin page title. --brion 16:56, 27 August 2007 (UTC)

Yeah, I know. I definitely am not advocating breaking all of the existing uses of percent-encoded links, I just prefer to have everything neat and tidy. Such is life. Mike Dillon 02:01, 28 August 2007 (UTC)

Yes, please report such things to bugzilla in future. :) Fixed. --brion 14:56, 26 August 2007 (UTC)

Thanks. PrimeHunter 16:02, 26 August 2007 (UTC)

Grey-zone Web-Spider Question

It's my understanding that using a web-spider such as wget is frowned upon because it can place a large load on the server and it is more efficient to download the entire database. My application does not warrant pulling down the database. I'd be interested in pulling 50 articles and each article that those 50 articles link to. I'm guessing that this would be on the order of 10-20K articles, well-below the size of the entire Wikipedia. This would actually make for a easy-to-implement but relatively powerful category suggestion tool that I would be happy to elaborate on if it's genesis were tolerated.

Therefore, would it be tolerated if I used wget or a similar spider under the condition that I ran it at a rate of only one page a second? Would this be broadly tolerated? It seems like all of the anti-spider notes are about people trying to download all of Wikipedia at 50 pages/sec. Chrisgagne 16:14, 3 September 2007 (UTC)

I think wget cares about robots.txt so it sounds fine to me, but I am no Wikipedia maintainer so I have no idea what their opinion is. Fetching 50 pages manually which is very feasible should put a much greater strain on the servers. Jeltz talk 22:20, 3 September 2007 (UTC)

Best would be for you to use the API, which you can access here. Prodego ^talk 22:30, 3 September 2007 (UTC)

If you just need a list of pages and the pages to which those pages link to, this can be accomplished quite simply using the API's prop=links. If you use this method, building a list of 10-20k articles should not be a tremendous load on the servers. Please, however, be sure you know your way around with wget quite well, specify a descriptive user agent (indicating how you can be contacted if there are problems), use a reasonable throttle (one request per second should be satisfactory), optimize your queries to reduce load, and most importantly, contact wikitech-l to notify and request permission from the server admins to begin running the crawler, lest it wind up blocked :). AmiDaniel (talk) 06:35, 4 September 2007 (UTC)

De-style links

In order to standardize Wikipedia articles, is there a way that I can "temporarily" select all and re-color the text so it will not "tempt" me to click on the supplied link(s) so just concentrate on the article itself. I can achive this by "highlighting" but is there a better way? Thanks. BRDaedaleus 23:33, 27 August 2007 (UTC)

Try this in your monobook.js. --Splarka (rant) 07:35, 28 August 2007 (UTC)

addOnloadHook(function() { addPortletLink('p-cactions','javascript:unstyleLinks()','Hide links','ca-hidelinks','remove styling from links'); });

function unstyleLinks() {
  var unstyleLinksCss = document.createElement('style');
  unstyleLinksCss.setAttribute('type','text/css');
  unstyleLinksCss.appendChild(document.createTextNode('#bodyContent a {color:black; text-decoration: none;}\n#ca-hidelinks {display:none !important;}'));
  document.getElementsByTagName('head')[0].appendChild(unstyleLinksCss);
}

Some modifications:

Can switch it on/off
Compatible with IE (orig. version doesn't work with IE6, not sure about IE7)

∴ Alex Smotrov 14:52, 28 August 2007 (UTC)

var unstyleElement = null, unstyleDisabled = true
addOnloadHook(function() { addPortletLink('p-cactions','javascript:unstyleLinks2()','Hide links','ca-hidelinks','remove styling from links') })
function addStyle(text){
 var st = document.createElement('style')
 st.setAttribute('type', 'text/css')
 if (st.styleSheet) st.styleSheet.cssText = text //IE
 else st.appendChild(document.createTextNode(text))
 document.getElementsByTagName('head')[0].appendChild(st)
 return st
}
function unstyleLinks2() {
 if (!unstyleElement) unstyleElement = addStyle('#bodyContent a {color:black; text-decoration: none}')
 unstyleElement.disabled = unstyleDisabled = !unstyleDisabled 
}

Dah-da dah da, dah da, dah da, daaaaaaaa.... --Splarka (rant) 00:53, 29 August 2007 (UTC)

Magic Word for My Preferences Language

Currently there's only the {{CONTENTLANGUAGE}}, but it's not completely satisfying. I was thinking about adding something like [{{fullurl:commons:Image:{{PAGENAME}}|uselang={{CONTENTLANGUAGE}}}} description page there] at MediaWiki:sharedupload. The flaw is, that if people have a prefered language, it gets overwritten, even if it's also set at Commons. Dose anyone have the ability to make a new MagicWord like {{CONTENTSETLANGUAGE}} or something similar. --Steinn inn 19:08, 20 August 2007 (UTC)

I believe a {{USERLANGUAGE}} has been propsed and rejected for the same reason {{USERNAME}} was: it would break cache, and be useless for non-logged-in-users (which make up the vast the majority of page views). You should be able to do what you describe with a little javascript though, using the wgUserLanguage and wgContentLanguage variables. --Splarka (rant) 07:32, 21 August 2007 (UTC)

A magic word returning the user language would not be a caching problem; parser cache objects are varied according to a number of factors, one of which includes the user language. 86.148.49.28 18:14, 21 August 2007 (UTC)

It would, however, cause problems when used in templates and conditionals, and result in effective corruption of link tables, such that templates and other things don't update properly. Whoops. 86.148.49.28 18:22, 21 August 2007 (UTC)

I'm not going to get into JavaScript. Thanks tough. If someone that knows JavaScript thinks it's a good idea, please help. I just don't know JS. --Steinn inn 12:11, 21 August 2007 (UTC)

BetaWiki uses a MagicWord called {{UILANGCODE}}. I don't know how Nike dose it, but it works. I dosn't look like he's using js. Maybe this could be put into MediaWiki. --Steinn inn 02:31, 29 August 2007 (UTC)

JS solution ?

Here is something that might work for you:

if(wgUserLanguage!=wgContentLanguage) addOnloadHook(uselangPrefOverride)
function uselangPrefOverride() {
  var upo = getElementsByClassName(document,'*','uselang-pref');
  if(upo.length==0) return;
  var r = new RegExp('uselang\=' + wgContentLanguage);

  for(var i=0;i<upo.length;i++) {
    var a = upo[i].getElementsByTagName('a');
    for(var j=0;j<a.length;j++) {
      if(a[j].href.search(r)!=-1) {
        a[j].href = a[j].href.replace(r,'uselang=' + wgUserLanguage);
        a[j].title = a[j].title.replace(r,'uselang=' + wgUserLanguage);
        for(var k=0;k<a[j].childNodes.length;k++) {
          if(a[j].childNodes[k].nodeType==3) {
            a[j].childNodes[k].nodeValue = a[j].childNodes[k].nodeValue.replace(r,'uselang=' + wgUserLanguage);
          }
        }
      }
    }
  }
}

What this does (note: not tested for all browsers):

If the userlanguage is not the same as the content language (which requires you to be logged in and have your default language chosen as something other than site default), it adds a load event.

It then searches the document for all objects of class="userlang-pref". Divs, spans, tables, wikitables, etc.

It then iterates over all these searching for any types of links.

It then iterates over all these links in the href="" field, searching for uselang=xx where xx is the content language of the site.

Any that match this criteria get uselang=xx replaced with uselang=yy where yy is the user's language choice. It also does a shallow (first match only) regex replace on the title="" and first level text nodes (if you have tags inside the link, the innderHTML is not modified on these).

<span class="uselang-pref">
<a href="http://foo?uselang=xx" title="http://foo?uselang=xx">http://foo?uselang=xx</a>
</span>

Becomes:

<span class="uselang-pref">
<a href="http://foo?uselang=yy" title="http://foo?uselang=yy">http://foo?uselang=yy</a>
</span>

The simple instructions are: add the JS above to MediaWiki:Common.js and then wrap links to be changed thusly: <span class="uselang-pref">[{{fullurl:commons:Image:{{PAGENAME}}|uselang={{CONTENTLANGUAGE}}}} link title]</span>

--Splarka (rant) 23:34, 23 August 2007 (UTC)

Automated upload of article edits

I noticed after some searching-around and FAQ-reading that WP (apparently) does not have an API. I'd like to know what is the acceptable method (if any) of automating article edits and article creation. I'd also like to know the same answer for Wikibooks. There seems to be plenty of guidance for people who want to download from WP, but not much for people who do not do very well with editing in the TEXTAREA of a web-browser. dr.ef.tymac 18:02, 28 August 2007 (UTC)

User:KeithTyler/mwpush.pl? —MC 18:23, 28 August 2007 (UTC)

There is an api: m:API. But the editing feature is not written yet. There are third-party libraries to automate uploads. And there is also the ability to use an external editor to edit WP articles. — Carl (CBM · talk) 18:27, 28 August 2007 (UTC)

Also see tools in WP:BOT, Wikipedia:Browser notes#Textarea_tools. (SEWilco 20:05, 28 August 2007 (UTC))

How can I disable the pop-up boxes?

You added it yourself by editing your monobook.js file. To remove it you'll need to remove the top four lines of text from that page. --Cherry blossom tree 08:17, 28 August 2007 (UTC)

Thank you! I'll do that now. :) - Jeeny ^Talk 08:21, 28 August 2007 (UTC)

Clear your browser cache. —Ilmari Karonen (talk) 09:58, 28 August 2007 (UTC)

About the Timestamp

Instead of 04:09, 28 August 2007 (UTC), can it be modified to this?

9:08 PM, 27 August 2007 (MST)

Note: When settings are updated on the user's preferences. -- PNiddy Go! 04:09, 28 August 2007 (UTC)

How would that work? Either each user gives timestamps according to their own personal time zone (which would be very confusing), or timestamps can't be treated as plain text (which would cause the developers a lot of problems). -Amarkov moo! 04:13, 28 August 2007 (UTC)

Here, try this in your Special:Mypage/monobook.js:

var autoLocalTimeStamps = false; 
importScript('User:Splarka/timestamplocalizer.js');

It runs User:Splarka/timestamplocalizer.js. It adds a button to your actions which, when clicked, will iterate over all the text nodes looking for timestamps. It should work on the default signature timestamps. The output is a bit spammy (raw output of the JS date .toString() function in your local time). Example of what it does on your timestamp for my timezone:

04:13, 28 August 2007 (UTC)
Mon Aug 27 2007 21:13:00 GMT-0700 (Pacific Daylight Time)

Note that I avoided innerHTML so this is a bit slow. You can have it run on every non-edit page load by setting false to true in the first parameter (be sure to test it first). Browser-screwups and bug reports ~~cheerfully ignored~~ welcome. --Splarka (rant) 10:46, 28 August 2007 (UTC)

Centralization of reference content

I'm not sure how to best explain this so bare with me. I almost always name my ref tags. It is certainly possible that when someone is editing an article, they could remove the first example of a named reference (the one with all the information). As such, would it be possible to create some sort of way to have all the named references displayed in one place and then add them in throughout the article. That way - if someone removes the "first example", the rest are not broken. Juan Miguel Fangio| ►Chat 04:47, 26 August 2007 (UTC)

I think someone else suggested that very recently. Have a trawl through the recent archives. I don't know what the response was or whether it would be easy to implement, but it is quite a good idea. Adrian M. H. 11:02, 26 August 2007 (UTC)

Thanks for the compliment, I did not find anything (probably because i didn't look proparly). Is anyone here familiar with what was said about this? Juan Miguel Fangio| ►Chat 03:42, 27 August 2007 (UTC)

Try looking through "See also" section of WP:FOOT. And its Talk page. (SEWilco 04:22, 27 August 2007 (UTC))

Yeah, I had looked through those previously and didn't see anything. Any thoughts? Juan Miguel Fangio| ►Chat 18:07, 27 August 2007 (UTC)
I have spent some more time looking through this stuff and don't see where it is possible to do what I've suggested. Is there a more appropriate talk page for this discussion? Juan Miguel Fangio| ►Chat 20:43, 28 August 2007 (UTC)

Sloppy workaround: Unless and until you find the previous discussion you were looking for, you might be able to use this *very ugly* workaround. Put the named references in a hidden "div" tag, and then just use them throughout the article.

WARNING: This is probably not a good idea for many different reasons, so I will *not* recommend that you use it in "live" WP articles, nevertheless, you can see what I am talking about by viewing this sample demonstration (note: this is a temporary link, subject to change or deletion without notice). dr.ef.tymac 20:58, 28 August 2007 (UTC)

Italics workaround?

In most style guides, foreign-language words should usually be italicized. The exception is when they occur in text that's already set in italics, when they should be in Roman type.

That's usually easy enough to do with Wikitext, but a problem arises our commonly used hatnote templates. See the one at the top of Raccoon, for example. The name of the genus Procyon should be in Roman type there because the hatnote is in italics, formatted by {{otheruses4}} -- but there's no way to do it.

This was unexpected behavior. A naive textual substitution would preserve the ''...'' markup within the template parameter; and that as a result italics would be turned off at the first '' (closing out the pair) and back on again at the second.

On examination I see this behavior might result from the definition of class dablink in MediaWiki:Common.css, which enforces italics. (This means that the '' markup in Template:dablink is redundant.) So is there a workaround I'm not seeing that would still allow the use of the template? Or is this even worth worrying about? TCC (talk) (contribs) 12:56, 23 August 2007 (UTC)

If it weren't for the redundant '' in the template, adding .dablink i { font-style: normal; } to MediaWiki:Common.css ought to fix it. I'm not sure what that '' is there for; if there's no prticular reason for it, we could just get rid of it. —Ilmari Karonen (talk) 16:18, 23 August 2007 (UTC)

Probably a leftover from before the class was defined. TCC (talk) (contribs) 10:58, 24 August 2007 (UTC)

I've removed it now. Assuming no problems come up, I'll make the change to Common.css tomorrow. —Ilmari Karonen (talk) 21:55, 27 August 2007 (UTC)

I've made the change. After clearing your cache (or waiting a while), you should now be able to see ''...'' in dablinks as roman type as expected. —Ilmari Karonen (talk) 10:30, 28 August 2007 (UTC)

Annotated Images

Can anyone point me in the direction of information on image annotation? I was looking at an article that contains an annotated image and don't know anything about how they are coded or how to move them within an article. None of the WP:Images-related pages that I've looked at have any information at all on image annotation. Thanks!! -Sarfa 17:37, 5 September 2007 (UTC)

Invisible characters around Norsk interwiki links

There are some invisible Unicode characters before and after the sidebar interwiki links for "Norsk (bokmål)" and "Norsk (nynorsk)". What are these for? --- RockMFR 05:41, 30 August 2007 (UTC)

They are U+202A and U+202C, directional formatting codes that mark the inclosed text as left-to-right. I can't see how they should be necessary, since other strings don't have them either and the Unicode bidirectionality system ought to ensure correct left-to-right display even if the string was to be displayed in an overall right-to-left environment. Don't know where these are stored though. Fut.Perf. ☼ 06:10, 30 August 2007 (UTC)

Wait a moment, on second thought: if these strings are stored centrally for all wikipedia projects (I'm just guessing), then there might be a need for directionality codes because these names contain brackets. They'd be necessary to ensure that the strings are displayed correctly in right-to-left wikipedias, such as the Arabic one. Otherwise they'd come out as "(nynorsk) Norsk" or something like that. Fut.Perf. ☼ 06:35, 30 August 2007 (UTC)

Ahh, that's probably it. --- RockMFR 06:37, 30 August 2007 (UTC)

Stub non-transclusion

Apologies if I use the incorrect terminology here - correct me please!

An example of what I'm attempting to do is:

Alraune (1928 film) includes {{horror-film-stub}} but doesn't need to be transcluded into Category:Horror film stubs as it (and similar articles) are categorized correctly into Category:Pre-1960 horror film stubs. The reason for this is that there are less than 20 films that would go into a "1920s horror film stub" or "1910s horror film stub", so no reason to create the tiny stubs, the overall one is fine. However, when doing "clean-up", these appear in a category that is to sub-categorized, even though there is no sub-category for these articles via a specific stub.

So, the base question is, is there a way to get the few articles with {{horror-film-stub}} and are included in Category:Pre-1960 horror film stubs not to show up in Category:Horror film stubs?

Thanks in advance for any assistance / advice! SkierRMH 22:38, 29 August 2007 (UTC)

Edit {{horror-film-stub}} to wrap the category link in a variable, something like this: {{{category|[[Category:Horror film stubs<noinclude>|*</noinclude>]]}}}. Then you can override the category in each page by providing the parameter, e.g. {{horror-film-stub|category=[[Category:Pre-1960 horror film stubs]]}}. Anomie 22:53, 29 August 2007 (UTC)

Thanks much!! Worked like a charm! Now to check the rest of the "by year" categories in film for stragglers... SkierRMH 05:08, 30 August 2007 (UTC)

Database hiccup?

The title explains everything. Got error message about 7 minutes ago. ~user:orngjce223 how am I typing? 16:47, 29 August 2007 (UTC)

Yep. --brion 20:26, 29 August 2007 (UTC)

Vandalism

This image; http://en.wikipedia.org/wiki/Image:Adam.jpg was replaced by someone in an act of vandalism, but sice I don't have a Wikipedia account I can't upload the proper image myself. I'm not sure if this is the right place to report this, but I've been trying to find the right place for half an hour now, and this is the best place I've found. —Preceding unsigned comment added by 68.222.240.200 (talk) 11:19, August 29, 2007 (UTC)

I've reverted the image. The best place for inexperienced users to get help is the help desk. Graham87 12:04, 29 August 2007 (UTC)

Actually, the best place for inexperienced users is NCH! Adrian M. H. 14:14, 29 August 2007 (UTC)

Tidy update

Any possibility of getting the bug below fixed? I believe the latest version of Tidy would do the trick (see Bug 9737 for more info). Cheers. --MZMcBride 20:40, 28 August 2007 (UTC)

one
two
three

Pity it doesn't skip 3. It would be a Monty Python sketch waiting to happen. TCC (talk) (contribs) 23:05, 28 August 2007 (UTC)

I don't see any rendering problem whatsoever. Has this been fixed? Nihiltres^(t.l) 00:45, 29 August 2007 (UTC)

Classic skin shows 1 one; blank line; 3 two; 4 three. Code is # one; # two; # three -- SGBailey 08:32, 29 August 2007 (UTC)

Template:E

Fixed. —Ilmari Karonen (talk) 10:32, 28 August 2007 (UTC)

Thanks -- SGBailey 08:37, 29 August 2007 (UTC)

This is wierd.

How come on the article diccussion of "Adam Powell", Instead of signing you username, it gives your IP? —Preceding unsigned comment added by 207.177.111.152 (talk) 01:06, 31 August 2007 (UTC)

Maybe because you haven't signed in? KTC 01:16, 31 August 2007 (UTC)

HughesNet issues

I have HughesNet Satellite Internet. For some reason, I can only stay logged on to Wikipedia if I use Internet Explorer. How can I get this Wikipedia to work again on Firefox, and don't suggest just using IE or using the secure server: IE freezes up on me and the secure server is slower and I can't use any of my tools on it. Psych less 21:30, 30 August 2007 (UTC)

Please ensure that you are accepting cookies from en.wikipedia.org. Goto Edit / Preferences, and then select the "Privacy" tab. Make sure that "Accept cookies from sites" is checked, and than wikipedia is not listed under exceptions. Also double check that you have opted to keep cookies "until they expire" and not "Until I close Firefox" or similar. AmiDaniel (talk) 21:43, 30 August 2007 (UTC)

Already done that, still have a problem. 69.19.14.34 21:44, 30 August 2007 (UTC)

Read advice at the top of this page. Conscious 06:27, 31 August 2007 (UTC)

In particular, you may need to use the alternative slower connection https://secure.wikimedia.org/wikipedia/en/w/index.php?title=Special:Userlogin (this is explained at the top of the page), which helps to avoid login problems on certain satellite ISPs. --ais523 17:24, 31 August 2007 (U T C)

Age of account for page move & semi-protected page edit

I would like a clarification / definitive answer on the exact length of time a new account need to be before they can 1) move page 2) edit semi-protected page. Help:Merging_and_moving_pages says "at least 4 days old." & WP:PROT says "disables ... fewer than five days old". However, at least quite a few experienced editors have suggested they are both 4 days. Would be nice to clear it up. KTC 22:56, 30 August 2007 (UTC)

Both the pagemove permission and the ability to edit sprotected pages are assigned based on the autoconfirmed flag, which is given to registered users whose accounts are 4 days old. It's the same wait for both features.--69.118.235.97 23:15, 30 August 2007 (UTC)

Cool, I've found the edit on WP:PROT that changed it which was vandalism from a sock. Corrected the info. on the page. KTC 23:25, 30 August 2007 (UTC)

Image description pages - file history

Recently, the file history at image description pages was converted to a table. From that point, image size is displayed in this table in kilobytes/megabytes. I suggest that it be reverted to the previous format (in bytes, exact size), because it's an easy way to check if two images are different. Conscious 11:15, 30 August 2007 (UTC)

Bobby Lane's Curse Seems To Extend to Apostrophe Usage...

Look here at the photo caption. You'll see an extra apostrophe at the end. When you click edit the page it's all fine, and not tags. I found one solution, that shouldn't work: click edit the page on that and you'll see that it should be putting the entire infobox in italics. User:Seraphimblade found a second way to make it work, but that shouldn't be working either. Anyone know the problem? -WarthogDemon 00:30, 30 August 2007 (UTC)

The Caption parameter is output surrounded by '' marks, to make it italicized; the combination of those '' marks and the '' marks around Times result in a '''', which corresponds to bold (''') and an apostrophe. You need to insert something in between the two sets of '' marks to get it to parse correctly; for example, <nowiki></nowiki> at the end of the Caption value takes care of it. Anomie 02:11, 30 August 2007 (UTC)

Or even better, <nowiki/>... a very useful wikitext element not many people are aware of. Gracenotes^T § 08:39, 30 August 2007 (UTC)

Monobook not working

Nevermind, the problem seems to have gone away... Dar-Ape 18:17, 30 August 2007 (UTC)

Uploading images

I used to be able to upload images with no problem. Now there is some kind of wizard that says my source file is empty. Before, the source file name was the image file name. Now that does not seem to work. What do they mean when they say Source file? Also, the second box says Destination file. Is that the image name or the article name or what? Thanks! --Mattisse 19:36, 27 August 2007 (UTC)

If you would prefer to avoid the new wizard-style form, Special:Upload still exists, but both systems use the Browse function (or require you to enter a filename) to find the file. Adrian M. H. 20:24, 27 August 2007 (UTC)

The file names I used to use do not work anymore. I have uploaded many images the old way. I don't know what the heck is the problem now. I'm giving up. Not worth the hassle. Thanks, --Mattisse 00:53, 28 August 2007 (UTC)

Can you give an example of a filename, and which browser (and what version) you use? It'll be easier to find and fix the problem if we can reproduce the difficulty. --brion 03:42, 28 August 2007 (UTC)

I am using Firefox 2.0.0.6. The file name is yarlang_tsangpo.png. It is a NASA image. Thanks! Mattisse 13:51, 28 August 2007 (UTC)

Do you have a particular suggestion for this situation? --Mattisse 19:25, 28 August 2007 (UTC)

That's a start; now can you describe exactly the steps you follow and exactly what error messages you get back? --brion 20:26, 28 August 2007 (UTC)

I put the image name in the first box. The second box confused me. I tried various things like the article name, the image name etc. Next I filled out the next box with the url of where I got it, that it was a US Government image, and that it was going to the article Yarlung Tsangpo River (Tibet). (The wrong NASA image is in the article now.) It is a NASA picture so I got that part right. The error message was always that the image was "empty". It was not recognizing my image, yarlang_tsangpo.png. In the past I have never had this problrm. --Mattisse 21:11, 28 August 2007 (UTC)

(unindent) Do you know what I am doing wrong? --Mattisse 00:50, 29 August 2007 (UTC)

This may be a change in Firefox; I have the same version as yours, and, as you report, I can't get it to upload files by entering just the file name. I suspect it may be looking for the file in a different directory, but if so, I can't figure out what directory that might be; it's not my home directory, nor my Desktop, nor the directory Firefox was started from, nor the one that is first shown when I click "Browse".

Anyway, entering the full path of the file works, as does clicking the "Browse..." button next to the box and selecting the file that way. I suggest you do the latter; it's by far the easiest way.

(Ps. The "Destination file" box is for the name the file will have on Wikipedia; this doesn't necessarily have to be the same as the name it has on your computer, hence the second box.) —Ilmari Karonen (talk) 14:02, 30 August 2007 (UTC)

==Heading 2==" always show this behaviour, but a single horizontal line ("----") does not. - Erik Baas 19:26, 29 August 2007 (UTC)

So is there any way in which we can integrate the horizontal rule code into the section headings to get the same result? Max Naylor 19:08, 30 August 2007 (UTC)

Section headings and infoboxes

No, I‘m not looking to solve the bunched up links problem, for example in Culture of Iceland the underscore of the section headings goes straight through the image, although behind it (you might have to have a wide screen resolution to make it work). I have posted a screenshot here [6]. I hope this makes things clearer. I don’t think it’s a browser problem, however if it helps I am using Safari on Mac OS X (I have looked at the same page in Firefox and the problem still persists). Max Naylor 12:50, 29 August 2007 (UTC)

Take a look at nl:Gebruiker:Erik Baas/Test; it demonstrates that section headings like "=Heading 1=" and

Occasional blank page at RMS Titanic

See WP:AN/I#RMS Titanic. Basically, sometimes the page is blank, sometimes it's not. x42bn6 Talk Mess 07:12, 31 August 2007 (UTC)

Must be the same as #Article not displaying properly. Conscious 07:33, 31 August 2007 (UTC)

In the case of RMS Titanic, the problem was fixed by replacing the DegMinSec coordinates in {{geolinks-US-cityscale}} with decimals equivalents. I have no idea why. Thatcher131 13:08, 31 August 2007 (UTC)

Please confirm that the issue is resolved when you put the template back, otherwise I'll need to track it down and see if there's a second bug with {{#expr}} in addition to the one I fixed a few hours ago. --brion 18:55, 31 August 2007 (UTC)

Article not displaying properly

When I go the the article Stuttgart, I get the article name, and the edit/discussion/history/watch etc tabs, however the content of the article does not display. If I click on edit, I get the edit screen, and I can also see the talk page. I have not had this problem with any other pages. DuncanHill 19:34, 30 August 2007 (UTC)

It looks to be a problem with Template:Infobox German Location although I'm not sure what it is. 69.202.63.165 19:42, 30 August 2007 (UTC)

Fixed. It had to do with having a <ref> tag in a parameter of Template:Infobox German Location. Cheers. --MZMcBride 19:50, 30 August 2007 (UTC)

Marvellous, fast and friendly, thank you! DuncanHill 19:53, 30 August 2007 (UTC)

MZ, are positive it was the ref tag? It seems to have fix itself when you reverted. It is very odd to say the least. —MJCdetroit 19:59, 30 August 2007 (UTC)

Indeed, this is still a problem on other pages that transclude this particular template (see Aachen, for example). 69.202.63.165 21:00, 30 August 2007 (UTC)

If you find a page that is still broken, purge the page's cache (click "edit this page" then change "edit" in the URL to "purge"). It looks like some other template was broken, but I don't know which one yet. --- RockMFR 21:14, 30 August 2007 (UTC)

I didn't realize I got edit-conflicted when I walked away after trying to post a follow-up. Purging solved the problem, though. --MZMcBride 21:15, 30 August 2007 (UTC)

Hmm... something's still wrong. I've been purging Göttingen repeatedly - sometimes it is blank, and sometimes it is normal. --- RockMFR 21:34, 30 August 2007 (UTC)

Both Aachen and Göttingen appear to work fine for me. DuncanHill 22:06, 30 August 2007 (UTC)

Something is definitely still wrong. Wikipedia:Help desk#Contents disappeared is discussing Mannheim. I have seen it change between blank and normal many times in the last 20 minutes. PrimeHunter 22:08, 30 August 2007 (UTC)

I think I got it. It's the en dash in the postal codes in the template call. The parser is inconsistently choking on it. Sometimes it throws an error, sometimes it just dies. This is one of the pieces of code causing the problem:

{{#expr:{{{PLZ}}}}}

It's weird that this is suddenly a problem. However, the templates should not be using the en dash at all if the template is going to be using expr on the param. They should all be using normal hyphens. I'm going to temporarily remove this hack from the template until the calls are fixed or the software can handle en dashes. --- RockMFR 22:35, 30 August 2007 (UTC)

Reported at bugzilla:11128. --- RockMFR 23:33, 30 August 2007 (UTC)

I don't understand why suddenly this is acting up. The expr fields that were in place for the postal and area codes have been functioning perfectly for ages (i put them there in the first place). Since it has suddenly stopped working, this is perhaps caused by changes to the expr functions at a more global level (there are a few undesirable features of the expr functions). Either way, thanks for fixing it up so that it works again. - 52 Pickup 07:25, 31 August 2007 (UTC)

This should be fixed since a few hours ago. (Purge as necessary.) The problem is an interaction between a bug in the ParserFunctions #expr function and a change in behavior between PHP 5.1 and 5.2, which is running on our newest batch of servers. (We're in the process of a major service upgrade, with a bunch of new servers in on our updated configuration system.) When an unexpected non-ASCII character cropped up, the error message could end up including an invalid UTF-8 sequence; later processing would abort on this character, blanking the page unexpectedly. I've worked around it so that particular problem shouldn't crop up anymore, but it's possible that other similar problems exist. Let us know if you see them and I'll track them down... --brion 19:06, 31 August 2007 (UTC)

UI for anonymous users trying to create a new page

When I click on a red link with the action=edit parameter passed in, I'm directed to an edit form, with active "Save page", "Show preview", and "Show changes" buttons. The only clue that I'm unable to create a new page is that the active tab is "view source". However, shouldn't the page be clearer and not feature any edit form whatsoever? Otherwise, it's misleading anonymous users into thinking that they'll be able to post new pages. There should probably be just a note similar to the current one given to anonymous users while trying to log in. Something along the lines of "Because you are not currently logged in, you cannot create new pages". The edit form should not show up at all for anonymous users. 69.202.63.165 19:15, 30 August 2007 (UTC)

It's a recent bug, the page creation form hasn't been visible for anons since they first outlawed anon page creation years ago. The page shouldn't be visible to anons at all, should just appear as a protected page. The problem is, bugs that only affect anons tend to go unnoticed by the developers, since they'd have no reason to edit logged out.--69.118.235.97 19:24, 30 August 2007 (UTC)

Oh okay. Thanks. Is this on Bugzilla somewhere? 69.202.63.165 19:28, 30 August 2007 (UTC)

Hard to say, I imagine it will show up there once enough people start to notice it.--69.118.235.97 19:38, 30 August 2007 (UTC)

Indeed, this is a bug; I'll try to see if I can figure out what is causing it in a bit. In the future, please submit these to Bugzilla, where they will be addressed far more quickly. AmiDaniel (talk) 21:47, 30 August 2007 (UTC)

I've filed as bugzilla:11140. --brion 19:12, 31 August 2007 (UTC)

Ok we seem to have worked that one out. Whee! --brion 21:43, 31 August 2007 (UTC)

Whitelisting of tinyurl.com

Tinyurl and similar url redirecting services are blacklisted on meta because they can be used to circumvent the spam blacklist. However, this also disallow linking the main site tinyurl.com from its article TinyURL. This could be done by whitelisting only the url http://tinyurl.com on MediaWiki:Spam-whitelist, that is, whitelist only the main url without anything following. So far, I considered adding tinyurl.com$ or tinyurl.com[^/], but I wouldn't like to experiment with the whitelist. I have also seen \b used in the meta blacklist. Any suggestion? Tizio 12:20, 29 August 2007 (UTC)

A further note: my reading of the files in [7], and in particular of the filter function in [8] is: first, the article is parsed and its external links are extracted and joined by newlines. Then, all urls matching http://+[a-z0-9_\-.]* followed by one line of the whitelist are removed from this string. If I am reading things correctly, tinyurl.com\b should work, as this regular expression only matches if the string tinyurl.com is at the end of the word (in this case, of the external link). Could someone with some more knowledge of PHP confirm this? Tizio 13:41, 29 August 2007 (UTC)

I was unable to get tinyurl\.com$ or tinyurl\.com\b or any of several alternatives to do what you want. They either allow all these or disallow them all. —Preceding unsigned comment added by Splarka (talk • contribs) 08:02, August 30, 2007 (UTC)

Thanks! I finally succedeed with www.tinyurl.com(?![/:]). Apparently, there is a bug with the software, so that the url that is reported in the spam message is the first url that matches the blacklist, even if it is included in the whitelist. Tizio 11:40, 30 August 2007 (UTC)

Reported bug: bugzilla:11121. Tizio 11:48, 30 August 2007 (UTC)

Heh, no wonder I couldn't get it working. Bah. --Splarka (rant) 07:19, 31 August 2007 (UTC)

Special:Log/delete

Is there any way to filter image deletions from the deletion log?--VectorPotential^Talk 18:20, 7 September 2007 (UTC)

I would say no, there is no way. Of course, you could filter them directly on the page using JavaScript ∴ Alex Smotrov 04:12, 8 September 2007 (UTC)

Some javascript that can do that (and more). --Splarka (rant) 08:46, 8 September 2007 (UTC)

Donation lettering overwrites long article titles

The new location of the donation text overwrites long article titles, at least in Firefox on Windows. Screenshot:

Tempshill 20:08, 1 September 2007 (UTC)

Double geographical coördinates

Dodge City, Kansas, like most US place articles, has its latitude and longitude given in a Geography section and in the external links, both of which naturally cause the coördinates also to appear in the top right corner of the page, like where an FA star goes. However, Dodge City's coördinates in the top corner are different and overlapping. I can't fix it; could someone please help? Nyttend 13:24, 1 September 2007 (UTC)

Help with tables

I can't figure out how to move the userboxes to the right so the userboxes aren't sitting on top of the userpage. More precisely, I want the userpage description on the left and the userboxes on the right next to it. See User:Hdt83/Sandbox. Can anyone help? --Hdt83 ^Chat 09:56, 1 September 2007 (UTC)

I think you should ditch all the deprecated table tags and use styled divs instead. Try to minimise table usage and when you need to use them, just use the wikitables. View the code of my userpage if it helps. Adrian M. H. 10:07, 1 September 2007 (UTC)

How hard would it be...

Hey, How hard would it be to make some script that meant when a page was added into a category, or an edit takes place on a page (on the user has specified, not all pages :P) A message comes up very similar to the 'You have new messages' box that comes up when someone edits your talk page? Thanks, Tiddly-Tom 22:12, 2 September 2007 (UTC)

I've thought the very same thing for a long time, only to finally reason that it wouldn't be that good an idea. The problem would be the ownership of articles could significantly increase, with edit wars likely to increase. Maybe we could do something for article discussion pages, but I don't know how useful that would be. violet/riga (t) 22:16, 2 September 2007 (UTC)

I will give you a little more information about what I was specifically thinking about. I am a member of WikiProject BBC and have assessed alot of articles. I wished to make a template (already made, but not implemented) which would put the page into a CAT that requested re assessment. I was wondering weather it would be possible to create what I mentioned in my first post to let me know when people want some re assessing :P Thanks, Tiddly-Tom 22:38, 2 September 2007 (UTC)

The first part of your request is covered by this script. See WP:US/R for script requests. Adrian M. H. 22:39, 2 September 2007 (UTC)

The second part is, I think, possibly covered by User:ais523/watchlistnotifier.js, which installs the same way as any other script, and lets you know when a page on your watchlist changes. From experience, I've found that small and grey is a lot less annoying as a notification than large and orange when it comes up almost all the time! --ais523 10:44, 3 September 2007 (U T C)

the [show] and [hide] links

How do I make the templates always be showing, instead of having to click on the [show] link to see them. I'd rather be able to click [hide] if I don't want to see the content.

I added ".NavContent{ display: block ! important; }" (without the quotes of course) to my monobook.css will allow some to show, but apparently not those with "class=autocollapse", which I've been told will perhaps need a JavaScript solution. Does such a script exist? Thanks -- Matthew Edwards 01:22, 2 September 2007 (UTC)

How many templates or tables are you working with? Wikipedia:NavFrame. I know of no scripts that can bypass that, though it should be possible. Adrian M. H. 01:30, 2 September 2007 (UTC)

I'm not working with any. It's just the ones that are at the bottom of articles that are hidden. I'd prefer to see them rather than having to click [show].-- Matthew Edwards 02:21, 2 September 2007 (UTC)

In practice I have noticed that if I turn off JavaScript in my Safari browser it makes collapse boxes show up in their expanded form. For example the boxes used to collapse each issue at Wikipedia:Conflict_of_interest/Noticeboard/Archive_1.

This seems consistent with the comment at Wikipedia:NavFrame that "Internet Explorer 5 and browsers which do not support JavaScript will render the elements without the [hide/show] links and will not collapse them." EdJohnston 22:44, 2 September 2007 (UTC)

Way to deal with legit links that are blacklisted

I thought of this. Some articles may have an ext link that gets blacklisted because of spam, and therefore has to be removed. How about creating a page which designates that certain links can only be added to certain pages without being blacklisted. Doesn't the site have a similiar page that prevents explicit images from being added to all but certain pages?--Avant Guard 21:57, 1 September 2007 (UTC)

Currently this only exists at the language/project level: MediaWiki:Spam-whitelist. Alternatively, such links can simply be included on the page in <nowiki> tags as unclickable but copyable text. --Splarka (rant) 07:10, 2 September 2007 (UTC)

What links here

Is it possible to have an option that shows what pages redirect or transclude to a page? - BANG ! 20:31, 1 September 2007 (UTC)

You can use http://en.wikipedia.org/w/api.php?action=query&list=embeddedin&eilimit=500&eititle=Main_Page to get a list of pages that transclude a given page, where you replace "Main_Page" in the URL with the name of the page you want to query. I don't think there's a way to just show what redirects to a particular page, aside from going through the entire what-links-here. Tra (Talk) 21:43, 1 September 2007 (UTC)

If you use my WhatLinksHere enhancement script, you can load up to 5000 links, then show only the redirects. — Randall Bart Talk 19:32, 2 September 2007 (UTC)

Also, What Redirects Here on Toolserver (might be down right now, then check it later) ∴ Alex Smotrov 20:08, 2 September 2007 (UTC)

Firefox 2 printing SVG graphics :background prints black

Every SVG image I have tried to print from a Wikipedia page allways prints as black - even though it displays correctly on the screen (and in print preview) Please help: you are using SVG increasingly and I just can't print the images that accompany the page. tyroneh@mailzone.co.za —Preceding unsigned comment added by 155.239.82.120 (talk) 12:55, 1 September 2007 (UTC)

This is a bug in Firefox on some platforms; see bugzilla:4968. --brion 18:29, 2 September 2007 (UTC)

overlap weirdness??

(overlap weirdness??)

(Still) getting some odd overlap issues.--69.118.235.97 22:44, 30 August 2007 (UTC)

Uploading images - 2

I asked a question about regarding this problem. I used to be able to upload images with no problem. Now there is some kind of wizard that says my source file is empty. Before, the source file name was the image file name. Now that does not seem to work. What do they mean when they say Source file is empty if I have added the image name? Also, the second box says Destination file. Is that the image name or the article name or what?

I am using Firefox 2.0.0.6. The file name is yarlang_tsangpo.png. It is a NASA image.

I put the image name in the first box. The second box confused me. I tried various things like the article name, the image name etc. Next I filled out the next box with the url of where I got it, that it was a US Government image, and that it was going to the article Yarlung Tsangpo River (Tibet). (The wrong NASA image is in the article now.) It is a NASA picture so I got that part right. The error message is always that the image is "empty". It is not recognizing my image, yarlang_tsangpo.png. In the past I have never had this problem. I would really appreciate help with this. Thanks! --Mattisse 13:18, 29 August 2007 (UTC)

Ok, really nothing much is different from how it's been for a year or two. There's just more annoying text at the top of the screen. :)

Once you have selected your file with the 'browse' button, don't change what's in the "source filename" box. If you change it, your upload will fail because Firefox will be looking for a different file than the one you selected, and it's probably not a file that exists on your computer.

(I can confirm that blanking or changing this box in Firefox causes the upload to fail with the message that "The file you uploaded seems to be empty. This might be due to a typo in the file name. Please check whether you really want to upload this file."It's also possible to get an "empty file" error sometimes if you have a very large file or some other kind of error, but it sounds from your description like changing the source filename after selecting it was the problem.)

You probably also don't need to change what's in the 'destination filename' box either. That gives you the opportunity to choose a different name for the file that you're uploading than the name that you have on disk -- but you don't have to change it. Just leave it as is unless you're sure you want to change it.

Once you've put in the information you want in the summary box and the license selector, just hit upload and it should work ok this time. --brion 15:50, 29 August 2007 (UTC)

Thanks! The file is 82 kilobytes. I don't change the name from its name on my desktop, except I remove the path to my desktop.

^^^^ that is the thing I told you not to do, don't do that. --brion 20:25, 29 August 2007 (UTC)

(I downloaded it from NASA but this is the way I have always done it.) Something must be different. I will try it again. Thanks! --Mattisse 18:54, 29 August 2007 (UTC)

I am frustrated out of my tree. I have tried everything. I am desperate! The website with the image is http://eosweb.larc.nasa.gov/HPDOCS/misr/misr_html/yarlang_tsangpo.html and the image is at http://photojournal.jpl.nasa.gov/browse/PIA03708.jpg The article I want it to go in is Yarlung Tsangpo River (Tibet) (Right now it has the wrong NASA image -- the canyon instead of the river.) Would you be ever so kind and upload it? I am naming the image yarlang_tsangpo_river_tibet.jpg. Help! --Mattisse 19:34, 29 August 2007 (UTC)

Please follow my directions. --brion 20:25, 29 August 2007 (UTC)

Ok, Ok I did. You were right. Thanks! (I'm not intentionally trying to be dumb.) Thank you! --Mattisse 22:55, 29 August 2007 (UTC)

Great! :) When we do the next overhaul of the upload form system we'll see if we can make it harder to make that mistake... --brion 18:32, 2 September 2007 (UTC)

Watchlists down?

My watchlist says that "Due to high database server lag, changes newer than 1268 seconds might not be shown in this list.". Ow! That's about 20 minutes.. Technical problems? Secretlondon 12:34, 6 September 2007 (UTC)

I can reproduce, 1769 secs here. Navou ^banter 12:43, 6 September 2007 (UTC)

My last entry 12:04 UTC (still). The only change is the lag figure, which keeps going up. Secretlondon 12:57, 6 September 2007 (UTC)

It looks like at 12.15 UTC the wiki went down. I'm guessing whatever runs the watchlists isn't up again. Secretlondon 13:03, 6 September 2007 (UTC)

Nothing useful to contribute, other than the fact that it's still down at over an hour of "database lag". (ESkog)^(Talk) 13:17, 6 September 2007 (UTC)

Watchlists are overrated anyway, and an undisputed source of stress. I predict a drop in edit and social conflicts for the duration. ←Ben^B4 13:20, 6 September 2007 (UTC)

Heh... we can hope. BTW, it's now at 4654 seconds (pushing 78 minutes). --Ali'i 13:32, 6 September 2007 (UTC)

Well I'm finding I can't check all the bureaucratic pages I need to.. Secretlondon 13:34, 6 September 2007 (UTC)

Is there any way the lag warning can give a figure in minutes rather than seconds? 5484 seconds is effectively meaningless (reminds me of the roadworks warnings that say "delays for 104 weeks, instead of the 2 years they really mean!). DuncanHill 13:45, 6 September 2007 (UTC)

How many more tasks should the servers perform due to the servers being overloaded? (SEWilco 17:52, 7 September 2007 (UTC))

Would it be more work to give an output in minutes rather than seconds?DuncanHill 12:04, 10 September 2007 (UTC)

When I clicked the blue triangle in front of an item to see additional changes, I was redirected to an empty page called "RCI0". For the next item it was "RCI1", then "RCI2", "RCM2", "RCL2" etc. I use FF 2.0.0.6. --Eleassar ^{my talk} 15:36, 6 September 2007 (UTC)

I've never heard of blue triangles, so you might want to take that up with one of the seven (!) extensions you have in your monobook.js. ←Ben^B4 15:39, 6 September 2007 (UTC)

No, it's a setting in your preferences, section "recent changes", tick "enhanced recent changes". Lupo 15:50, 6 September 2007 (UTC)

I get blue triangles in recent changes, but not my watchlist. ←Ben^B4 16:13, 6 September 2007 (UTC)

They work now. And yes, there appear to have been technical problems. WP was inaccessible for a few minutes, then in a database lock as the slaves caught up to the master. At a wild guess, one of the servers reset. >Radiant< 16:17, 6 September 2007 (UTC)
- This might be relevant: Wikimedia projects downtime Thu Sept 6, 12:00 UTC; it's at about the time the problems were reported. --ais523 16:25, 6 September 2007 (U T C)

Ad Ben et al: section "watchlist", tick "Expand watchlist to show all applicable changes". It's still disfunctional. --Eleassar ^{my talk}

Mine's fine. I'll go check again. --Pupster21 Talk To Me 12:06, 10 September 2007 (UTC)

Oh shoot mine says this. "Below are the last 29 changes, as of 12:06, 10 September 2007." I think something's wrong. --Pupster21 Talk To Me 12:07, 10 September 2007 (UTC)

That sounds right. Ryan Postlethwaite 12:09, 10 September 2007 (UTC)

I just realised that the watchlist time is GMT I'm US EDT. Still, I'm lagging one minute at least. --Pupster21 Talk To Me 12:13, 10 September 2007 (UTC)

Tips and reminders somewhere on the page

I was thinking that it might be a good idea to add reminders somewhere on the page (think of those texts on a loading screens of some video games), the reminders can range from simple things like "You can edit this page by clicking edit this page on the top of the page" and "If you see anything wrong in this article be bold and change it" to specific reminders that are only given out to members of a specific project.

There is also the issue of where to put these reminders, they could be on top or the bottom, etc., and when they're visible, they could be visible all the time, only while you're editing, you could opt out, or it could become visible when clicking a help me, I'm a newbie button, etc.

And if we wanted to get really fancy, we could have a targeted reminder based on your contributions, e.g. if this is your first visit to Wikipedia it would say something simple like "You can edit this page by clicking edit this page on the top of the page", but then if you're advanced and really good at editing long articles then it could give you tips on editing long articles and provide links to long articles for you to skim down.

Don't get distracted by the last part about targeted reminders, I just thought the normal reminders might make people more aware of those Wikipedia features that they wouldn't usually hear about until they really really needed it.

I originally posted something like this here and they said gave it a medium to cold welcome, but now I moved it here and added some of their improvements. Jeffrey.Kleykamp 21:25, 3 September 2007 (UTC)

I personally feel that there is far too much UI clutter as it is and that this would simply add to it. Such a "tip of the day" tool would probably be best done through use of a simple system message, such as that displayed above/below the edit box, or on MediaWiki:Watchdetails, as opposed to as part of the software itself. However, in both of those forms, I really think it will be more of a nuisance than an aide. AmiDaniel (talk) 06:19, 4 September 2007 (UTC)

Contributions from an IP range

Are there any tools out there to help monitor contributions from a range of IP addresses? I see that VandalProof has a "Contribs from IP Range" feature. I also see that a temporary bug fix (T3035) that allowed range searches was implemented then reverted. So are there any other options? This obviously would be a helpful alternative to range blocking, having the ability to monitor ranges of IP addresses where vandals or other disruptive editors with dynamic IPs are known to have originated. Obviously, feel free to point me in the direction of past discussions if I have missed something already discussed in my past searches. Thanks.-Andrew c ^[talk] 21:07, 3 September 2007 (UTC)

--> WikiScanner. -- Agathoclea 21:22, 3 September 2007 (UTC)

Yeah, I tried wikiscanner out, but the database they are using is a month old, so it doesn't really help with vandalism fighting, except for looking into the distant past to dig up old, missed edits. But thanks for the reply!-Andrew c ^[talk] 21:34, 3 September 2007 (UTC)

Spouting about my own software, but ... WP:VANDALPROOF (now unsupported) has some basic IP range contrib-tracking tools, but they aren't the fastest in the world. m:WikiMonitor will allow you to monitor edits from specific IP ranges in real-time; however, I don't believe I have the preloading of range contribs implemented at present time (project has been set on the backburner for the time being). Both are fairly bloated Windows apps (run fine under Wine though), but will get you about what you need. AmiDaniel (talk) 06:23, 4 September 2007 (UTC)

Grey-zone Web-Spider Question

Best would be for you to use the API, which you can access here. Prodego ^talk 22:30, 3 September 2007 (UTC)

Proposal for better recent changes

The people watching the recent changes are doing a great job, and there exist many good tools available, but regretfully some vandalism still slips through.

This probably came up before, but is it possible to add better vandal-fighting tools to the recent changes itself (so implemented by the Wikipedia server itself, rather than third party tools)? One thing I can think of is having a "checked" flag for the diffs that were examined by a logged in user who is different than the original contributor of that diff.

Such a thing is obviously not perfect, but it would at least give an idea of the changes that slip through without any examination whatsoever.

This can put a lot of extra strain on the database, maybe, but I'd argue it would be worth it. Oleg Alexandrov (talk) 17:08, 2 September 2007 (UTC)

Such thing already exists (see Help:Patrolled edits) and is built into Mediawiki since version 1.4. It is however apparently disabled on Wikipedia. There's also Extension:Patroller, which somewhat extends the functionality. And yes, I can imagine it being a significant extra server strain on a wiki this large. Миша 13 18:13, 2 September 2007 (UTC)

From my experience with it on Meta, it's quite useless because people are too lazy to click on [mark as patrolled] every diff they view and wait for "Marked as patrolled" page to load. I think it needs to be implemented in JS just like watchlisting. MaxSem 18:15, 2 September 2007 (UTC)

I believe you mean AJAX, and the relative bug would be bug:11002. Unfortunately, patrol has far greater problems than just this, and I doubt that it would really be effective in combating vandalism. My take on the matter has always been that, while server-side mechanisms may help somewhat, the best implementation of a countervandalism utility is client-side. I've had a few attempts at client-side patrol-like tools, such as WP:VP2; however, these have been quite unsuccessful in large part because of their inability to gain and maintain an active and large userbase. Similar server-side attempts have had the same problem. AmiDaniel (talk) 06:05, 4 September 2007 (UTC)

Related changes

A question on Related changes on categories. Currently the talk pages of articles under wikiprojects are in the wikiprojects articles category. For example the Wikipedia:WikiProject Yorkshire articles are to be found in Category:WikiProject Yorkshire articles. You can use the related changes function to get the changes related to any of the talk pages in the project. But is there any way to get the changes relating to the actual articles associated with these talk pages? Obvoiusly you do not want to put project related categories on the actual article which would be one way to do it. Keith D 19:28, 3 September 2007 (UTC)

Chinese/Japanese language options

I am appaled that there is no possible way to view Wikipedia in Japanese/Chinese text! Would you please fix this problem SOON?!--65.100.245.114 19:19, 3 September 2007 (UTC)

The Chinese Wikipedia can be found at http://zh.wikipedia.org and the Japanese at http://ja.wikipedia.org . Both can be accessed using the In other languages box at the left of the main page. -- lucasbfr ^talk 20:03, 3 September 2007 (UTC)

"Pages transcluded onto the current version of this page" doesn't show all transclusions

Hi. While tracking down Double geographical coördinates, above (which is probably conflicting/redundant coordinate templates), I noticed that not all transcluded templates are displayed in the list of transcluded pages. For instance, when editing {{Geobox Settlement}}, {{Geobox label}} which appears in the code is not in the list. Is this new/intended? Thanks, Saintrain 18:17, 3 September 2007 (UTC).

{{Geobox label}} is in {{Geobox Settlement}}, but under an #if. When the "category" param is not specified, {{Geobox label}} isn't expanded, and doesn't show in Pages transcluded. — Randall Bart Talk 18:49, 3 September 2007 (UTC)

A similar phenomenon is discussed at Wikipedia:Template limits#Conditional inclusion. There are times when it's important for nontranscluded templates not to count as if they were transcluded. — Carl (CBM · talk) 18:58, 3 September 2007 (UTC)

So, since there were no params (since the template wasn't called during the edit) to trigger the condition, the template wasn't transcluded. Makes sense for template counting but not very handy for debugging these interminably nested templates. Submitted as bug 11172. Thanks, Saintrain 21:24, 3 September 2007 (UTC).

Wikipedia:Tutorial/Tabs

Wikipedia:Tutorial/Tabs seems to have been vandalized, but I don't know where the modification's source is so I can't fix it. Can somebody help here? ColourBurst 15:08, 3 September 2007 (UTC)

Reverted. It was embedded in another page and transcluded over. x42bn6 Talk Mess 15:10, 3 September 2007 (UTC)

Notice to Wikipedia:Researching with Wikipedia

I don't know what is causing it, but while visiting WP for the first time on this computer (as an anon), I had an horrible mixup of site notices, "Your continued donations keep Wikipedia running!" • • "Learn more about using Wikipedia for research" •. Which are at the same spot on the screen (FX and IE). I couldn't find the page causing this in the mediawiki namespace, but I confirmed it on an other browser. The links disappear after visiting Wikipedia:Researching with Wikipedia. -- lucasbfr ^talk 10:19, 3 September 2007 (UTC)

That was in an older version of the anonnotice changes that have been happening recently, as far as I can tell; that problem should be avoided with the current version. --ais523 16:42, 3 September 2007 (U T C)

By the way, I believe the site anonnotice is now being hidden in MediaWiki:Common.js. Cheers. --MZMcBride 16:55, 3 September 2007 (UTC)

Maps with dots on

For example, in Guildford, there is a map of Surrey in the Infobox with a red dot. From perusing the template, I have determined that the red dot is positioned using the longitude and latitude coordinates given in the infobox. What I haven't sussed is how the uploaded map knows what its long/lat are so that the calculation of dot position can be done. What does one do if the dot is in the wrong place after all this? -- SGBailey 20:41, 2 September 2007 (UTC)

The dot is added by some CSS, I think, which overlays an image of a dot on top of the image of the map (if you hover the dot and the rest of the image, you'll see a different image description page is linked from the dot than is linked from the rest of the image). If the dot's in the wrong place, it would probably be a problem with the template that does the overlaying. --ais523 10:48, 3 September 2007 (U T C)

Yes. BUT the map on which the dot is placed needs to have the lat/long of its corners defined so that the CSS can work out where to put the infobox specified lat/long for the dot. Where/How do the map's corners (or centre and scale or ... ) get defined please. -- SGBailey 13:12, 3 September 2007 (UTC)

At Template:Location_map#Available_maps, there is a list of templates for different countries. If you go into the edit page for each of these templates, this is where the map boundaries are defined. Tra (Talk) 13:30, 3 September 2007 (UTC)

Lag times

Is there a page where we can check the current server lag situation? When I am working on disambiguation, I fix a template, then go check the WLH again. Sometimes all the transcluding pages disappear immediately, sometimes they disappear in bunches over the next few minutes, and sometimes it's an hour or more before they move. This is independent of the lag in updating the page itself. When I go to a page and see the old version of the template, I know that's server lag, and I know I can action=purge to make it update. Unfortunately, even action=purge on the transcluding page will not cause the WLH to update.

So anyway, rather than reloading the page every 10 seconds until doomsday. I'd like a way to know that I should move on to something else and come back later. — Randall Bart Talk 23:24, 31 August 2007 (UTC)

Check the job queue by going to special:statistics. The bigger it is, the longer it will take for links in the templates to be updated. Since I edit on off-peak hours the longest wait I've had for template links to update is an hour. Graham87 09:34, 1 September 2007 (UTC)

Thanx, I'll check that number when I have lag issues. I was hoping to see more than one number, however, because I have empirical evidence that there's a WLH lag and a Redirect lag that that can hit when other taks are proceeding quickly.— Randall Bart Talk 18:37, 3 September 2007 (UTC)

Colored vandalism warnings

At User talk:212.219.90.77, and on at least one similar vandalism-dominated talk page I remember, when a user is blocked the message is colored. But when he's unblocked and lower-level vandalism warnings resume, they are also colored, even though the same template used on most pages shows as white. That is probably a bug. Art LaPella 18:21, 10 September 2007 (UTC)

I'm afraid I really don't follow -- can you try to explain the problem again? AmiDaniel (talk) 04:50, 11 September 2007 (UTC)

I understand, and it was a result of the (then used) template for being blocked, not having a </div> to close it properly. You can go in and manually place the tag to close it and it fixes the rest of the page, as you'll see was done to the page you mentioned [9]. Here is what it looked like before the closing tag was added, [10]. Hope that helps explain it! (The template has since been fixed, this issue should not happen now). ^Ariel♥Gold 04:53, 11 September 2007 (UTC)

Thank you. Art LaPella 20:31, 11 September 2007 (UTC)

Login Problem with Microsft Internet Explorer (MSFT IE) - HELP!

Microsft Internet Explorer (MSFT IE) version: 6.0.2900.2180.xpsp_sp2_gdr.070227-2254

I have a problem. I had not done any editing on Wikipedia for a while. But, when I tried to strart up again, around May, I could not get Wikipedia to open (come up) if I was logged in.

If I deleted my cookie and re-booted, I could go to Wikipedia fine, surf around and make edits on some articles. But, if I logged in, "nope", the progress bar just sits there. It just sits there, says: "Login Successful," but the progress bar is just stuck. If I use "Secure Login" it logs in, is very slow (I am on dial-up), and then logs me out when I go to a new page. It never opens the page, any Wikipedia page. Delete the cookie and re-boot, can go to any page. Login seems to go fine, says that you're logged in, but doesn't open any pages.

So, after a week or so of frustration, I started using a different account. That worked fine until today (2007-08-25). Now it is doing the same thing. Logged in, nothing just a stuck progress bar. Delete cookie and re-boot, view articles.

I tried the things on the help page. But, they don't seem to work either.

I don't have problems with any other site. Just this one. It has to be something that the sever admin's did, made a change in the cookie or something. 90% sure of that. I can connect fine, I can be connected for six, eight, tweleve hours at a time, and surf around to a hundred sites. Of course with dial-up I do avoid sites like U-Tube and the like. I can't believe I am the only one having this issue..???

I think it must have something do with COOKIES? I did a search for "WIKI" on the hard drive and found 26 files (2 *.HTM and 24 *.TXT). I deleted all of them, rebooted, searched, "none", signed on this time it worked, but then when I went to the next page...LOCKUP... So, I searched again:

C:\Temporary Internet Files
- Cookie:don@en.wikipedia.org/

C:\Documents and Settings\Don\Cookies
- don@commons.wikimedia[1].txt
- don@en.wikipedia[1].txt
- don@en.wikipedia[2].txt
- don@en.wikiquote[1].txt
- don@en.wikisource[1].txt
- don@www.mediawiki[2].txt

With just ONE sign in I got SEVEN new files. And notice the "1"'s and "2"'s....this must have something to do with it?? I had to delete them all to get to this page. I have never seen a sight that gives you so many cookies (unless they are ads embeded in a webpage) with just ONE signin. It would also seem that having a "1" and a "2" might cause a conflit?

Is anybody else going to step up to help? I see questions in this forum being answered that don't even belong here, but my problem goes mostly ignored. Thanks

IE 6.0.2900.2180.xpsp_sp2_gdr.070227-2254
Windows XP - Home Edition - Version 2002 - Service Pack 2
Earthlink Total Access - Dial-Up

What is the deal?

207.69.139.157 00:49, 7 September 2007 (UTC) (WikiDon)

Have you tried logging in from another computer? --MZMcBride 00:55, 7 September 2007 (UTC)

Hi. I can't help you solve it, but the cookies are normal. If you visited any wp sister sites, you have cookies. Many sites will give you multiple cookies, like google, which usually gives you three or more. Thanks. ~A H 1^(TCU) 00:59, 7 September 2007 (UTC)

Well, it HAS to be something with the way that MSFT IE caches passwords/user names. I downloaded Mozilla Firefox and it is working fine. Of course I have not checked the box "Remember me" on the Wikipedia Sign in (log in) page, or do I say "Remember" to "Do you want Firefox to remember this password?" pop-up either. Choosing "Not Now" instead.

So, if I can login on the same computer, all else being the same, and work fine, it MUST be the way that Passwords and/or Usernames are being cached.

How do I clean out the MSFT IE and/or Wikipedia caching of Usernames and/or Passwords, besides clicking "Delete Files" and "Delete Cookies"? They must be hidden in there somewhere.

Please give suggestions. 207.69.139.149 15:33, 11 September 2007 (UTC) (WikiDon)

Radio buttons not functioning in history pages

The radio buttons are no longer functioning in the history pages; the topmost radio button bullet can no longer be moved down, although the bottom one still works. I found this to be the case in both Internet Explorer and Firefox. This problem started yesterday. — Quin 23:17, 4 September 2007 (UTC)

Strange, right after I made this post it started functioning again. Server error, or maybe the devs are just fast? — Quin 23:24, 4 September 2007 (UTC)

This was a bug in wikibits.js (see above), that has now been fixed. You likely had it cached client-side and either explicitly or unknowing purged that cache after you made this edit. If you continue to have problems, please try Ctrl+F5 to perform a hard-refresh. AmiDaniel (talk) 02:56, 5 September 2007 (UTC)

$1, $2, and $3???

I've seen many scripts with $1 and $2 and I'm wondering what are these? Also, there is a flaw when I reverted in popups when It said "Reverted to revison $1 dated $2 by $3 using popups" and now, what do these mean? Does it mean for instance, something will set $1 and will later use $1? Like $1 will be set by input text and then used by an action? I don't understand :/ --Coaster geekperson 04 00:03, 3 September 2007 (UTC)

I don't know which script uses these but most likely it is a placeholder in JavaScript. The script then replaces $1, $2 and so on with extracted or computed text. It looks a lot tidier than, say, var s='Revert to revision '+var1+' dated '+var2+' by '+var 3+' using popups';, because JavaScript is a naturally messy language (to me, anyway). But there is a script that did have a flaw. Try bypassing your cache, to see if that helps. If not, ask the script creator. x42bn6 Talk Mess 15:17, 3 September 2007 (UTC)

I don't think it's JavaScript; it's more likely to be a server-side script like Perl or PHP, where variables of this sort are used. *Dan T.* 15:52, 3 September 2007 (UTC)

I think it might be the popups, though. It does seem to be a placeholder of some sort. I think you should ask Lupin about the script if this is the case. x42bn6 Talk Mess 16:01, 3 September 2007 (UTC)

But the script worked fine with some people, but not me, or others. Coaster geekperson 04 23:06, 4 September 2007 (UTC)

I notice that you're simply importing the most recent version of the script -- it is possible (though I haven't investigated it) that a change was made to the script which broke it and that has now been fixed. I wouldn't worry about it :) AmiDaniel (talk) 02:59, 5 September 2007 (UTC)

Bot needed

Hello,

We need someone to write and maintain a bot that will police the use of the controversial Template:Unreferenced.

This template, currently on over 10,000 articles (Maybe it's 100,000; I don't know; I clicked "next 500" until I tired of it), says, "This article does not cite any references or sources." The problem is that this is usually untrue, and the tag has therefore lost credibility. There is steadfast opposition to changing this statement to say "sufficient" instead of "any'; other editors enjoy pointing out that there are other template tags such as Template:Refimprove that complain about references with lesser wording than "Unreferenced", and that the "Unreferenced" tag is needed, despite the widespread inaccuracy.

It occurred to me that a bot could be written that would fix this problem. It would run through each article that is tagged with Template:Unreferenced, and if there are any single-bracket links in the article at all, the bot would assume it's a reference, and change the template to Template:Refimprove. This would strengthen the integrity of Template:Unreferenced.

Any volunteers?

Thanks - Tempshill 22:01, 4 September 2007 (UTC)

You're looking for WP:BOTREQ. They'll be happy to help you there. Cheers. --MZMcBride 22:08, 4 September 2007 (UTC)

Will do, thanks. Tempshill 23:38, 4 September 2007 (UTC)

Auto throttling of articles for current events

Looking at the Steve Fossett article today, I had an idea for handling of current events. There's not point in having an article edited dozens of times in a few minutes. Therefore there should be throttling of the specific article, something like this:

When an article is edited more than 20 times in 20 minutes, the article automatically goes into throttle mode
While in throttle mode, registered users may not edit the article until five minutes have passed since the last edit
While in throttle mode, unregistered users will get a semi-protect message
When edits fall to less than five in the last hour, throttle mode ends

The end of throttle mode is deliberately hard to hit. An article could remain in throttle mode for days. I think this is a good thing. I thought about adding a filter to not auto throttle when all the edits came from one person, but I think occasionally inconveniencing those who don't use "show preview" is also a good thing. I realixe this would require coding in the MediaWiki software, but I would like to just toss the idea around here. — Randall Bart Talk 20:04, 4 September 2007 (UTC)

What happens if the last edit in throttle mode is blatant vandalism? what happens if there is breaking news that needs to be added that occured <5 minutes since last update? I think it is a poor idea. If things get bad enough, articles can be semi protected against unregistered users however I feel that such a function would add little value to the funcitonality of this software. Chrislk02 (Chris Kreider) 20:06, 4 September 2007 (UTC)

I don't see any value in slowing the pace of edits. What is the objective here? Tempshill 22:04, 4 September 2007 (UTC)

I don't like this idea either. One of the things that impresses people about Wikipedia is how fast it reacts to breaking news, often outpacing traditional news outlets. Also as Chris said there might be problems with accidentally fixing vandalism in place. the wub "?!" 20:16, 4 September 2007 (UTC)

I've checked the magic words page, and the section help page, and I haven't found any way to auto-hide the TOC. Is this possible to do? (With an essay, not with an article.) When the TOC is needed for navigation, but the length is visually distracting, it would be great to be able to have that hidden on default, and expanded on demand. Is there a way to do this? (I know you can limit them in LocalSettings.php, but I mean for global effect, so all users see it) Thanks in advance! ^Ariel♥Gold 18:45, 4 September 2007 (UTC)

One option might be to put

{| class="collapsible collapsed"
!Contents
|-
|
__Toc__
|}

Although that sort of thing generally tends to be limited to userspace pages. You might also want to put __Notoc__ and then manually choose only the main headings and list them separately. Tra (Talk) 19:12, 4 September 2007 (UTC)

Yeah right now it is in userspace. I thought I read that the multiple section TOC function isn't working? When you say manually choose the main headings and list separately, could you explain what you mean? I'm not really clear on that. Thanks! (I'll try the collapsing table for now) ^Ariel♥Gold 19:27, 4 September 2007 (UTC)

Well, that collapsed table works, sort of. The first click on "show" for me does nothing, but if I click it twice it works. That may be something related to my browser, and I guess not a big deal. I'm going to need to figure something else out, I think, because a TOC will really be needed eventually. Wish there was a way to get it to only see second-level headings, lol. ^Ariel♥Gold 19:32, 4 September 2007 (UTC)

Have you tried using {{TOClimit}}? Alternatively, try making subheadings you don't want to appear in the ToC using ;subheading rather than ===subheading===.-gadfium 19:55, 4 September 2007 (UTC)

I just put that template on a second ago. Tra (Talk) 19:58, 4 September 2007 (UTC)

Oh my gosh Tra, that just rocks! That's exactly what I want, thank you thank you! And thank you for adding it, that is just wonderfully helpful! And I think it would be appropriate if moved into the WP: space as well, so it is just perfect! I had previously had the headers as non-TOC headers, as gadfium suggested, but I preferred the style and look of the TOC type headers. Thank you again!

^Ariel♥Gold 20:20, 4 September 2007 (UTC)

diffcheck() in wikibits.js isn't working if you're logged out

Someone who knows how to work bugzilla, please reproduce and add this:

log out
visit an article with several items in its history
click the history tab
click on some of the radio buttons
observe that they aren't updating properly.

This is related to diffcheck() or one of its surrounding supporting functions historyRadios() or histrowinit() in http://en.wikipedia.org//skins-1.5/common/wikibits.js and I can't figure out how being logged out can affect it. ←Ben^B4 11:42, 4 September 2007 (UTC)

Fixed. --Gmaxwell 14:03, 4 September 2007 (UTC)

== Header == with expandable divs. AmiDaniel (talk) 06:13, 4 September 2007 (UTC)

We're not supposed to hide anything that's part of the article, only navboxes etc. ←Ben^B4 09:31, 4 September 2007 (UTC)

I think the user may have meant something automatic for headers on a page like this. --MZMcBride 09:43, 4 September 2007 (UTC)

Collapsible section heads

Can page rendering show those "+/-" expand/collapse/expand all/collapse all icons next to section headers?

This way on long articles, it gets much easier to see just the one section, see sections in context, or navigate without having to refer back to the contents box.

For serious users this would probably be a Good idea and improve usability a lot :) FT2 ^{(Talk | email)} 01:11, 4 September 2007 (UTC)

Sure, it can be done, but it's generally considered poor style. Simply wrap up the

Watchlist

A certain edit made to one of my watched pages isn't showing up in my watchlist, although it did a few hours ago. Sah msi dea Tel Aviv 23:18, 3 September 2007 (UTC)

Maybe the article in question was deleted, Jeffrey.Kleykamp 01:32, 4 September 2007 (UTC)

Can you point to which specific edit or article is missing? It's possible the page was deleted or moved, among other less-common causes. AmiDaniel (talk) 06:15, 4 September 2007 (UTC)

Flymeoutofhere's edit to my subpage, User:Sahmsidea/WikiProject Tel Aviv. Sah msi dea Tel Aviv 23:29, 4 September 2007 (UTC)

Oversized divs

Is there anything that can be done to prevent edits like this one and this one, where oversized divs disable the normal functioning of the page? In Firefox you cannot see the edit or history tabs, but in IE7 you can, and in Opera there is no effect at all. -- zzuuzz ^(talk) 19:02, 3 September 2007 (UTC)

The edits are easy to revert, and editors who persist in making them will end up blocked. It's not a big deal. — Carl (CBM · talk) 19:09, 3 September 2007 (UTC)

I run Firefox 2.0.0.6, and aside from making the page reaaaaaaaaaaaaaaaaaaly long, it has no effect on my ability to use the page. --Deskana (talky) 19:12, 3 September 2007 (UTC)

I am also running Firefox 2.0.0.6, and I get a page that is all black for the first 80%, then the bottom 20% is gray. The article tab is visible, but none of the other tabs or text are visible, and passing the mouse over where links ought to go doesn't get any signs of mouse over. But CBM's point is good. You can revert them when you find them, and users will learn or be blocked. — Randall Bart Talk 23:16, 3 September 2007 (UTC)

That's the same effect as I get. Any spammer who tries it will normally find themselves insta-blocked when I see it, but that's not really the point. It is quite difficult to revert for any editor who doesn't know how to access the page history, either through recent changes or directly in the URL. Otherwise, if you arrive at such a page normally, there is literally nothing you can do. It seems to me the software could and probably should be able to prevent it. -- zzuuzz ^(talk) 01:46, 4 September 2007 (UTC)

is there any valid reason why a user markup should need to be able to manually set a z-index? Maybe disable parameters such as z-index, or transparent text, that are far more useful to vandals and spammers than legitimate users? FT2 ^{(Talk | email)} 01:08, 4 September 2007 (UTC)

Such tricks are used by a number of bits such as some magic templates. IMHO they tend to be unreliable, though, as they often stomp over each other. :) So probably they all should get revamped some way. --63.135.227.17 18:55, 4 September 2007 (UTC)

External edit links

Would it be reasonable and/or technically possible to block requests to action=edit from a non-wikimedia referer? This is something that came up in regards to the (currently in arbcom) THF/Michael Moore incident, but IIRC it was expressed by several people that allowing external sites to directly post "edit some wikipedia article" links creates a wider danger. --Random832 19:40, 5 September 2007 (UTC)

Some browsers (and bots) don't use referers, so this is will have to taken into account if this is implemented. In addition, there are utility websites capable of forwarding an HTTP request with an altered or removed referer header; the malicious sites could merely use an indirect link. Semi-protection should work for the cases you mention, and otherwise, blocking would. I'm also not sure how effective this would be: people know how to find the "edit this page" button. Gracenotes^T § 20:33, 5 September 2007 (UTC)

To clarify; my proposal was to deny when A) there is a referer present and B) it is not from a wikimedia website - not for ones from any particular malicious website, or if there is no referer --Random832 22:53, 5 September 2007 (UTC)

Annotated Images

Wikipedia linked to url links automatically

Hello I recently installed new hard drive previously had Wikipedia link to url's as an option. I could have Wilipedia info on each site I visited. Can someone tell me how to obtain this feature again. Thank you. Can I be notified if someone makes a post response to my inquire? —Preceding unsigned comment added by Xandr (talk • contribs) 16:47, 5 September 2007 (UTC)

Discrimination article

The article Discrimination is having technical problems. The interlanguage links and categories are not being displayed. Please have a look and fix it. Thanks! - Gilliam 13:17, 5 September 2007 (UTC)

Done. Closed a HTML comment and copied the badly mangled interwikis (someone must have edited with a non-UTF-8 editor) from an old revision. Lupo 14:17, 5 September 2007 (UTC)

User page Header

I am trying very hard to design my user page even though html and wiki markup is completly new to me. My header seems to smush the words "Welcome to my user page....." to the left of the links and their icons. Can anyone figure out how to put these words centered and about the icons and links. Feel free to change it yourself or tell me how to do it. Thanks in advance! Monkeynoze 02:19, 5 September 2007 (UTC)

You want your the header above the icons or in the middle of a ring of icons? --MZMcBride 02:21, 5 September 2007 (UTC)

Thank you very much! This is exactly what I wanted!! Monkeynoze 10:40, 5 September 2007 (UTC)

Deletion related pref.

Is there a preference or the like I can set somewhere where I can chose not to have an auto-summary when deleting? It would be a realy great function if there was. Thanks. -- Anonymous Dissident^Talk 21:12, 4 September 2007 (UTC)

It should be possible via JavaScript; I know how to add a tab with the ability to do that, but I don't know how to modify the current tabs. --MZMcBride 22:23, 4 September 2007 (UTC)

Actually, I highly recommend installing the CSD AutoReason script. Very, very handy. EVula // talk // ☯ // 23:10, 4 September 2007 (UTC)

Here is the JS you seek. You may have to adjust the string matches depending on language/project:

if(wgAction=='delete') addOnloadHook(function() {
 var r = document.getElementById('wpReason');
 if(r.value.indexOf('content ')== 0 || r.value.indexOf('page was empty') == 0) r.value = ''
});

It is made to only remove the reason on document load, if it starts with "content " or "page was blank" (should get "content before blanking was: blab blah", "content was: blah blah" and "page was blank" reasons.

--Splarka (rant) 07:18, 5 September 2007 (UTC)

Bulleted list problem?

I have noticed that carriage returns in bulleted lists are currently being counted as separate lists when using templates. Below is an example.

London's Transport Museum Photographic Archive

In this list, the subitems (except the first) appears with two bullets (they weren't before). However, I can remove the superfluous bullet by removing the carriage returns:

London's Transport Museum Photographic Archive
- Passengers boarding a train at Covent Garden, 1921** Looking down station platform, 1921** Piccadilly line train at station platform, 1923** Station facade, 1925** Booking hall, 1928** Another view of booking hall, 1928** Lift with ticket inspector, 1936** Station facade, 1937

I noticed this in the External Links in the Covent Garden tube station article, which I fixed. Is this happening with anyone else? Andrew (My talk) 16:04, 4 September 2007 (UTC)

It always happens to me on Firefox, and never happens to me on IE. The carriage returns in such a list are supposed to split it into separate lists from the software's point of view. (It also happens with the other sorts of lists, including definition lists (:) which are used on talk pages, but the effect is only obvious with bulleted and numbered lists.)

Some more examples:

An indented definition list

using double spacing (the HTML is wrong, but it looks reasonably right)

A numbered list
using single spacing

but with a double space before this line

A bulleted list inside a definition list

with a carriage return between each line

- A bulleted list inside a bulleted list

- with a carriage return between each line

This is a 10-times-indented with colons comment

- - - - Indented 11 times with stars

Indented 12 times with colons

As you can see, there are all sorts of problems caused by these sorts of things; some show up in Firefox, and some show up in Internet Explorer. In some cases, the HTML is 'wrong' but all browsers I know of show it correctly anyway. --ais523 16:27, 4 September 2007 (U T C)

(If you're wondering; IE has a double-size break inside the bulleted list inside the bulleted list, and a really massive break before the 11-times-indented line; Firefox shows 11 bullets on the 11-times-indented line and 2 bullets on each of the bulleted-inside-bulleted lines, and both browsers go back to 1 for the third element of the numbered list.) --ais523 16:31, 4 September 2007 (U T C)

Fixed. The template you were using contained an extraneous linebreak which caused the parser to decide each bullet was a separate list.

As for those other problems User:ais523 mentioned, I've never had much trouble as long as a few points are remembered:

Don't put blank lines in the wikitext in the middle of the list.
To insert a linebreak in the middle of a numbered/bulleted list item, use <br /> (without a linebreak in the wikitext) to avoid starting a new list item.
Like this. Comments can be used to hide the wikitext-linebreak if you really want a wikitext-linebreak in there.
When nesting, repeat the preceding formatting. So for the "10 colons, 11 stars, 12 colons" example above, this should "work":

This is a 10-times-indented with colons comment

Indented 11 times with stars
Indented 12 times with colons

Anomie 16:57, 4 September 2007 (UTC)

Agreed. It's surprising how often I have to refactor discussion pages to avoid such problems, though... --ais523 17:04, 4 September 2007 (U T C)

Cheers you guys - it seems to be working normally now. Andrew (My talk) 14:10, 5 September 2007 (UTC)

WHOIS & DNSstuff

The Special:Contributions page for an anonymous user, identified by the user's numerical IP address, contains at the bottom a link labelled "WHOIS" (and some other links) that lead to dnsstuff.com (as, for example, on Special:Contributions/24.15.28.218). I was happy when that facility appeared not so long ago, because it gave a convenient way to find out where an anonymous user was posting from, but now these links stopped working for me; instead I get this message: This type of DNSstuff tool access requires DNSstuff membership. If you would like to do any of the following, please join us for just $3/mo. They offer a limited number of free lookups that would soon be exhausted. Could these links be replaced by links to a free site offering the same functionality? --Lambiam 20:11, 1 September 2007 (UTC)

The links work for me. Perhaps you have made quite a large number of lookups on them? I think any site that offers these kinds of services would need to put in some kind of throttle to stop people overloading their servers. Tra (Talk) 21:18, 1 September 2007 (UTC)

Just remove their cookies, everything will work as it was before. MaxSem 21:20, 1 September 2007 (UTC)

I quit the browser session, removed all cookies containing the string "dnsstuff", fired up my browser and tried the WHOIS link — alas, with the same message as before. Then I tried another browser that I rarely use, and never for this kind of stuff, again with the same (lack of) result. It looks as if they keep track of the querying IP address in some other way. When I access them through their home page, I can do a lookup; but the lookups with the IP address attached to the uri are blocked. --Lambiam 22:14, 1 September 2007 (UTC)

I have the same problem, but if I use a different computer, it will work for me :) Tiddly-Tom 22:14, 2 September 2007 (UTC)

The issue here seems to be that dnsstuff is blocking certain common ISPs on account of the volume of traffic that wikipedia is sending their way. From the perspective of dnsstuff, wikipedia may as well be launching a DOS attack against them based on the sheer amount of traffic we send their way. I'm slightly surprised their servers haven't started to crumble under the pressure we're putting on them.--VectorPotential^Talk 23:15, 5 September 2007 (UTC)

MediaWiki:Newsectionsummary

I'm somewhat annoyed by this new action. Since a few days, when you create a section, it will make an edit summary with an arrow linking to WP:AES rather than to the actual section. What are people's opinions on this? >Radiant< 09:51, 6 September 2007 (UTC)

The arrow links to WP:AES, but the actual name of the header links to the appropriate section, like a ToC box. I don't mind it. The back arrow's use is pretty consistent throughout the project. --MZMcBride 10:30, 6 September 2007 (UTC)

Well, normally when you edit a section, you get a forward arrow that links to a section. The difference between a forward and a back arrow isn't all that obvious in a list of links. >Radiant< 10:38, 6 September 2007 (UTC)

It's removed now, never mind. >Radiant< 15:56, 6 September 2007 (UTC)

format problem or bug?

The "see also" section in article Chick slaughtering is not appearing in the main article, but it appeared when it is in edit mode. The format appears correct, and I purge my servers and browsers cache, but that didn't work. Please check the articles talk page. Can somebody fix this formatting problem or bug?--PrestonH 03:01, 6 September 2007 (UTC)

There was an unterminated tag, in this case, a ref tag. An unterminated div can have the same effect. I've fixed it.-gadfium 03:27, 6 September 2007 (UTC)

That's what I get for not reading through my edits more closely. Thanks a lot =D Darkcraft 13:16, 6 September 2007 (UTC)

Automated AFD closer

I have added one to my .js, but it doesn't seem to be working. My mainly used browser is firefox. Can anyone recommend one that should work with FF? I would be very appreciative. -- Anonymous Dissident^Talk 10:41, 5 September 2007 (UTC)

No worries. I have got one :) -- Anonymous Dissident^Talk 10:18, 6 September 2007 (UTC)

New section

Hi, I have a couple questions.

It appears to me that recently, the software was changed so that, when a user adds a new section to a talk page, the edit summary will be "New section:section_name" instead of simply summarizing with the section name. Am I correct?
If yes, when was that decided and applied? And why so?
Is there a particular discussion on the topic where this was decided? If yes, where can I access it?

Thanks in advance for your answers! ^^ Zouavman Le Zouave 23:27, 4 September 2007 (UTC)

You're correct that it was a recent change to the software. The relevant system message page is MediaWiki:Newsectionsummary. A developer decided to implement the change; I'm unsure if there was consensus and I'm unable to find the specific revision update at the moment. Cheers. --MZMcBride 02:13, 5 September 2007 (UTC)

Developers make many changes to the site interface. The actual functioning of the Mediawiki software isn't an matter of consensus; the developers have a lot of discretion. But this change has a useful function: now there is a link to the new section, but before there wasn't. — Carl (CBM · talk) —Preceding unsigned comment added by CBM (talk • contribs) 02:47, 5 September 2007 (UTC)

Wasn't really any "discussion" about it, except for, I guess, this :D. It was introduced in rev:25445, and no objection was raised before it went live here. I would presume that the reason for this addition is that an edit summary should state what one is doing with the edit -- in this case, adding a new section -- whereas simply providing the title of the section could be confusing. If you would rather it be removed, either here or bugzilla: is likely the place to state why. AmiDaniel (talk) 02:54, 5 September 2007 (UTC)

I remember bringing this up somewhere (don't remember where) a while back. The previous versions of MW didn't distinguish between an edit comment the user just typed in and an automatically generated edit comment when the user started a new section. Also, I use the section link in edit comments heavily, so having a direct link to new sections is very useful. I suspect others will also find it useful. — Carl (CBM · talk) 12:06, 5 September 2007 (UTC)

It is certainly an excellent addition. Adrian M. H. 12:14, 5 September 2007 (UTC)

There are a couple of "bugs" with this: one is that if the section title includes a wikilink then the link to the new section will not work; the other is that if the new section is too long then the end of the edit summary will be cut off and again the link won't work. I'm not sure if either of these are fixable but it seems they should be fixed if possible. - MTC 16:21, 6 September 2007 (UTC)

yes, I saw that first bug yesterday; it's annoying when watching changes at RFF (where editors are asked to use links as titles). Adrian M. H. 16:27, 6 September 2007 (UTC)

Yes, I've also noticed this bug. [11] as against my contrib log. Frankly it's just an annoyance but the first couple of times I thought I'd miskeyed my edit summary. Pedro | Chat 19:50, 6 September 2007 (UTC)

You should file a bug at bugzilla.mediawiki.org about this. — Carl (CBM · talk) 20:08, 6 September 2007 (UTC)

Seems to be fixed in the repo (rev:25598 and later). AmiDaniel (talk) 20:37, 6 September 2007 (UTC)

Secret society of wikipedians

If a message on your talk page led you here, please be wary of who left it. The code below could contain malicious content capable of compromising your account; if your account appears to be compromised, it will be blocked. If you are unsure whether the code is safe, you can ask at the appropriate village pump.

what is this wikipedia secret society stuff ? I found this bot WP:TW and was interested in it. But it seems that most of the links in the info page end up in this. I should have trusted the conspiracy theorists.Jeroje 03:57, 31 August 2007 (UTC)

What? That is just standard boilerplate text asking you to not add any javascript code you may be unsure of. (This was added after a vandal went around and told users that they would lose their account unless they added a script from a certain off-site page, or something like that.) The Village Pump is just like the info line in your telephone... so I don't know what you're talking about. Titoxd^{(?!? - cool stuff)} 05:30, 31 August 2007 (UTC)

In order to install a script like twinkle, you need to edit Special:Mypage/monobook.js, a page which contains user-customizable javascript that will execute on every Wikipedia page view. That warning is there because a lot of people do not understand javascript and could be fooled into adding malicious code that could thieve your password, install assorted viruses on your home computer, etc. For the most part, any user script, such as WP:TW, that you find is harmless, however. AmiDaniel (talk) 18:28, 31 August 2007 (UTC)

Passwords are not made available in plain text to scripts, and JavaScript is not enabled on those pages which accept passwords, so the first assertion is wrong, and the second is broad, ambiguous and not quite as clear-cut as that, either. Please stop spreading misinformation. 86.134.25.176 21:21, 6 September 2007 (UTC)

Nevertheless it wouldn't be too hard to collect passwords from naive users through a carefully-crafted javascript routine. — Carl (CBM · talk) 21:27, 6 September 2007 (UTC)

There is a way for user scripts to create a legitimate-looking login form (with the correct URL) and monitor any passwords typed within. However, it would be best not to go into specifics. Basically, you shouldn't install any user scripts unless you can trust them. Tra (Talk) 21:49, 6 September 2007 (UTC)

Wikipedia:Village pump (technical)/Archive AU

Too much character lag

WP Size

Job queue

The Hebrew Wikis

Icons interfering with donations message...

Extension:CategoryTree

How can I disable the pop-up boxes?

Image sizes as percentages?

Please do this (new software changes)

Longstanding image/text space problem

Table/list with uneven borders

Why isn't this image working?

Is there an error in footnotes?

"This page is either protected or semi-protected."

Images in Faeroes

template show / hide (showing by default when should be hide)

Monobook not working

Section headings and infoboxes

Template:E

Help needed with complex template

gets the users’ list of a particular DataBase when name of the DB is supplied through query

Schadenfreude article first line displays funny

Watchlist stuck at 14:51 GMT?

Watchlist notification thingy

5.1 Speakers + Xbox 360 + HD/VGA Cable + PC Monitor??!?!?!?!?!?!

New donation links

Watchlist - Individual section watching & "Special Aritcles"

Template:08 ?

Is there a way to change wgContentLanguage?

If I use the "back" button on my browser, all Wikipedia's button's stop working!

Non-free software screenshot categorization

Twinkle problem

Does this page look weird, or is it just me?

MediaWiki:Accesskey-pt-preferences

What happened to image sizing?

Software update

Link names on page history/user contribs

overlap weirdness??

Metadata problem solved

"This is a User Page" technical/obfuscation question

"What links here"

What links here?

Using the "orphanBacklinksOnSpeedyDelete" feature on Twinkle

html to pdf

Notification of page moves

cite.php

Block log problems

Blocked text not working?

Image sizes as percentages?

Please do this (new software changes)

Redirects with + don't work

Grey-zone Web-Spider Question

De-style links

Magic Word for My Preferences Language

JS solution ?

Automated upload of article edits

How can I disable the pop-up boxes?

About the Timestamp

Centralization of reference content

Italics workaround?

Annotated Images

Invisible characters around Norsk interwiki links

Stub non-transclusion

Database hiccup?

Vandalism

Tidy update

Template:E

This is wierd.

HughesNet issues

Age of account for page move & semi-protected page edit

Image description pages - file history

Bobby Lane's Curse Seems To Extend to Apostrophe Usage...

Monobook not working

Uploading images

Section headings and infoboxes

Occasional blank page at RMS Titanic

Article not displaying properly

UI for anonymous users trying to create a new page

Whitelisting of tinyurl.com

Is there a way to change `wgContentLanguage`?