This page contains discussions that have been archived from Village pump (technical).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.
So a while ago (I think over a year, but I may be overstating things), I made a pitch to start auto-adding archive links to cite-web template links. That discussion got bogged down and I lost track of it, but it seems, at least from the numerous dead links I've seen, that the end result was apparently no action.
I tend to use archive.org, I don't care which archive source we use, I don't care if we use a random number generator to choose among a list of possible archive sites and invite anyone who wants to join that list if they can handle the requests.
We could do this via bot. All we need is a simple bot that goes to cite-web template links, checks if there are archived copies, if it's dead (or I would prefer even if it wasn't dead yet), add an archiveurl and archivedate parameter (if we can make it smart enough and the archive source offers multiple archived copies, we should aim for the copy closest to the access date (if available, otherwise closest to the date the citation was added, if available, otherwise latest), barring those with bad http response codes). I'm not an expert on Wikipedia bots and I also am not generally active enough to wade through the bot approval process, but someone could write this.
However, this is also something which ought to be doable on the Wikipeda software-level, and/or the Wikipedia foundation. I would just like to see some action by some one to combat this problem. Of course, I'm sure someone will ask, why not me? Maybe someday, but given my level of Wikipedia involvement, I'm not betting on soon. Jztinfinity (talk) 19:20, 5 February 2012 (UTC)
Random web pages archive OK, but very few periodicals allow themselves to be archived by free services, opting rather to use a pay-to-view archiving solution. An automated archiving solution would preferentially maintain links to the lowest quality sources, unfortunately. Not a bad idea to do, it's just a matter of looking at the consequences of putting such a solution in place. --User:Ceyockey (talk to me) 19:51, 5 February 2012 (UTC)
I think that one thing which could help to deal with link rot where a periodical is involved would be implementation of the findtpl extension. I found this a minute ago while poking around for ways to search citation template parameter values for a particular periodical title, The News Journal, which has a 'move to pay-to-read archive' policy in place which breaks all article links in a matter of weeks after use here. The use of the findtpl extension in conjunction with some bot work could provide a useful assist to people who want to monitor references to particular periodicals in citations. --User:Ceyockey (talk to me) 23:16, 5 February 2012 (UTC)
Renumbering template parameters
When adding a new parameter to a template that calls {{Infobox}}, it's necessary to increment the numbers of all the subsequent parameters. Does anyone have a script or tool to perform this tedious task? Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits20:37, 5 February 2012 (UTC)
Broken bullet points
I don't know if this is a new issue, but I've never encountered it before until twice recently: Bullet points and numbered lists failing to wrap around floating images (example) breaking away from their sentences. (I can't remember the other example!) There's a similar, not-so-critical problem with indents not working when placed to the right of an image. Can this be fixed with code or is it a software issue? nagualdesign (talk) 07:56, 29 January 2012 (UTC)
What browser/OS are you using? Your example looks OK to me using IE8 on Windows 7 (but I think I remember seeing the problem you have described using IE9 on Windows 7). DH85868993 (talk) 08:11, 29 January 2012 (UTC)
Thank you but that was not what I wanted, and moving the panorama down was unnecessary either way - it's the smaller image causing the interference. nagualdesign (talk) 08:32, 29 January 2012 (UTC)
I moved the panorama because it just looked odd at the top of the article. Feel free to move it back up if you so wish, but {{-}} is the key to fixing the text-wrapping. Goodvac (talk) 08:39, 29 January 2012 (UTC)
It's a small bug in the CSS implementation. The bullets/numbers are placed outside the list item content (list-style-position: outside;). This is fine when there is nothing in front of them, but interferes with floating content. What basically happens is that the list margins are ignored. Using inside would fix lists beside floating content, but will break list indenting. — Edokter (talk) — 11:17, 29 January 2012 (UTC)
Examples in IE today 29 Jan 2012
Has the bug always been there? The layout isn't ideal in IE8 or IE6 either (see screenshot). I don't recall seeing this before. - Pointillist (talk) 11:40, 29 January 2012 (UTC)
It's been there as long as I can remember (May 2009). I try to avoid placing left-aligned images low down in a section, in case they protrude into the next. Right-aligned images protruding into the next section are less of a problem. --Redrose64 (talk) 17:12, 29 January 2012 (UTC)
Thank you to those who have taken the time to follow this issue up. Whilst it is arguably very minor it's certainly worth adding to the To do list for the programmers. Internet Explorer with all of its quirks is the bugbear of many a website designer, but design for it we must. 'Favouring the other leg', so to speak, is not technically a solution. nagualdesign (talk) 11:20, 30 January 2012 (UTC)
It's not just IE; all browsers exhebit this behaviour. And as I explained above, there is no solution. It's basically a design consideration. — Edokter (talk) — 11:27, 30 January 2012 (UTC)
AFAICS it is just IE 9 (the top screenshot in the image above) where the list numbers and bullets are outside the layout box. Happens in both 32-bit and 64-bit IE9 on 64-bit Windows 7, using Vector, Monobook and Simple skins. This is a distinct bug using IE9/Win7. It doesn't happen in Chrome, FF or older IE: they just exhibit the known list-style-position: outside; issue described by Edokter. - Pointillist (talk) 10:10, 6 February 2012 (UTC)
Ah, didn't catch that. I'm curious to see what would happen with list-style-position: inside; in IE9 (which I don't have). In any case, I'm afraid we still would not be able to fix it. Best practice is to avoid lists next to left-floating elements. — Edokter (talk) — 11:53, 6 February 2012 (UTC)
Help moving list defined references
I want to move a lot of text with a lot of references from Game of Thrones (TV series) to Game of Thrones (season 1). Unfortunately, the source article uses list-defined references (WP:LDR). This means I must copy the entire huge list of references to the target article, and then I get cite errors in the reference lists in both articles because many footnotes defined in the list are now used in the other article. Is there some fix or workaround for not having to sort out manually every single reference? Or is there a tool that can convert the article back to normal inline references format, which allows much more flexible editing? Sandstein 20:12, 5 February 2012 (UTC)
I fixed the cite errors by converting the entire article (including the LDR portion that you added) back to the inline format using my References segregator script. These are the steps I performed:
I used regex to replace uses of the {{r}} template with plain ref tags.
I copied and removed (cut) the entire LDR list from the wikitext.
I clicked the "Segregate refs for editing" button.
I removed all the empty ref tags from the lower text box.
I pasted the LDR list into the lower text box (after its existing contents).
I saved the page, which merged the citation templates from the lower text box into the upper text box and removed the unused footnotes in the process.
I only did a quick check for obvious errors and duplicate references, so you probably should examine the list of references to see if there is anything that did not convert correctly. PleaseStand (talk) 20:54, 5 February 2012 (UTC)
Thanks for your prompt and competent help! I'm now trying to fix the corresponting errors in the source article using the same procedure. Do you remember the regex you used for the first step? Sandstein 21:03, 5 February 2012 (UTC)
Or are there other regex experts around? Simply using {{r|(.+)}} → <ref name="\1" /> doesn't work, as this will treat lines that contain multiple {{r}} entries as one match. (As an aside, this weird references format really annoys me; it's ridiculous that one needs regex knowledge and a special script just to move text from one article to the other.) Sandstein 21:34, 5 February 2012 (UTC)
I agree that {{r}} and {{rp}} are annoying. If your regex engine is sufficiently advanced, you could try {{r|(.+?)}} instead; otherwise you could try {{r|([^}]+)}}. See Regex#Lazy quantification for an explanation of how this works. Anomie⚔00:36, 6 February 2012 (UTC)
I have fixed the cite errors on the source article using the same procedure. I used the "Search and replace" feature of the enhanced editing toolbar (Search for: {{r\|([^}]+)}}, Replace with: <ref name="$1"/>, Treat search string as a regular expression) to deal with {{r}} (note the backslash that escapes the vertical bar). PleaseStand (talk) 02:38, 6 February 2012 (UTC)
All other observations aside, this is probably a very valid point. Anyone else tried editing with iOS or Android devices? ▫ JohnnyMrNinja23:24, 5 February 2012 (UTC)
general question - not responding to Nin — Would searching for a term not jump to that term location in the edit window, as it does in a regular browser instance? --User:Ceyockey (talk to me) 23:28, 5 February 2012 (UTC)
There is no search for term native for Safari/iPhone. I have an applet that I use for that purpose, and it will highlight terms, but not jump to them. PuppyOnTheRadio (talk) 03:35, 6 February 2012 (UTC)
Of course that only works for 3GS or later, as earlier models can't support iOS 5, if the users even want to update. Perhaps we should look at a simple addition to the editing window for these types of devices? ▫ JohnnyMrNinja07:36, 6 February 2012 (UTC)
Serious note - it's an issue with the browser and not the site. I have tried playing with different options to be able to adjust it and had no joy. Again, going to Help talk:Mobile access#Scroll down when editing a wiki page I've added another way of getting around this, which is to copy all text to notepad, make the edits, and tyhen copying it all back to the edit box. It's a pain to do but I have not been able to find any better work-around. PuppyOnTheRadio (talk) 08:38, 6 February 2012 (UTC)
Local image search?
I've been trying to search local images but the search results include Flickr images. Is there an easy way to filter these out? ▫ JohnnyMrNinja01:48, 6 February 2012 (UTC)
The problem is that Commons images display as local WP pages to search engines, so many images that come up are actually on Commons. ▫ JohnnyMrNinja06:33, 6 February 2012 (UTC)
That does it, thanks! Although, it still doesn't make sense that the default search for this wiki contains images hosted on another wiki. Makes maintenance a bit of a pain. ▫ JohnnyMrNinja07:03, 6 February 2012 (UTC)
Most image searchers probably want to view or use images without caring where they are so it makes sense to include Commons by default. But it would sometimes be helpful if the namespace selector had an option for local files. PrimeHunter (talk) 15:08, 6 February 2012 (UTC)
That's what I mean: if I click on a diff from my watchlist then I often want to go straight to the amended section to make a further edit; but instead I have to edit the entire page. I thought we used to be able to go straight to the section and edit it. DrKiernan (talk) 10:36, 6 February 2012 (UTC)
@Someguy1221: The old behaviour was that section edit links would only show in a diff if the more recent version of the two being compared was the current version; in that way you couldn't accidentally edit a section within an old version. The new behaviour is that some pages show section edit links in a diff: and when they are shown, they still follow the old rule about only showing when comparing to current version.
Editing a section from a past version is technically possible as a combination of oldid=id with action=edit§ion=n, but it would be unwise to call this from a diff or oldid because of the risk of wikitext corruption due to sections' renumbering. Incnis Mrsi (talk) 10:13, 6 February 2012 (UTC)
Yes, this used to be possible and was recently broken. You can still edit from a diff if you transclude a special page onto the page, though no one knows why. According to comments in the bug, this has been fixed in MediaWiki 1.19, which will hopefully be rolled out within the next few weeks. Ucucha (talk) 13:53, 6 February 2012 (UTC)
Yeah, it's hisioulsy anoying to me -- 98% of the time when I want to edit and it's not a rollback/undo, I'm looking at a diff, because after all, how else would I check what's new?. The change has been annoying enough that I've simply not even bothered a few times. ♫ Melodia Chaconne ♫ (talk) 14:54, 6 February 2012 (UTC)
Movepagetext not showing proper message to admins.
I've noticed that I'm not getting the admin-specific message when I move a page: "Note to admins: The "leave a redirect behind" option should only be unchecked when..". I left a note on this on the talk page here but there's been no response. Is this just me or is no sysop seeing the appropriate warning? Tassedethe (talk) 17:01, 6 February 2012 (UTC)
Ah, that explains that then. It does seem that the new message is missing much of the older one's functionality e.g the specific warnings about trying to move User pages. Tassedethe (talk) 18:46, 6 February 2012 (UTC)
System messages in German?
Please see this permanent link. At the bottom of the page an extra-large red font text says that there are <ref> tags but no <references /> tag in the article, but why does it say that in German? Maybe we've decided to gradually turn into a German-language Wikipedia, but there is one already... I'm confused :) --Theurgist (talk) 17:54, 6 February 2012 (UTC)
A database error has occurred. We apologise for any inconvenience this might have caused. The most likely cause of this problem is a search or other operation that took too long. Possible reasons include:
* A search where all words are in quotes. Try searching without the quotes initially or add a few more words outside the quotes to restrict the search;
* An exceptionally large personal watchlist (probably over 10,000 items); or
* Exceptionally heavy load on the database servers.
Technical details about this error:
Last attempted database query: (SQL query hidden) Function: SqlBagOStuff::set
MySQL error: 1637: Too many active concurrent transactions (10.0.6.50)
I keep on getting that too - only in the last 20 minutes or so, but I did get the same error a few times last week as well. It's not only when you try to save a page either - I got it the first time I tried to come here. SmartSE (talk) 19:14, 6 February 2012 (UTC)
I originally got the Database error as mentioned above. Now I see odd things with images that come from Commons. Over at Commons, it's very slow to load a page with multiple images, and not all the images load. If I pull up a WP page that has images from Commons, such as Museums in East Texas or Museums in West Texas, some of the images will load, but in some places it instead links directly to the WP page image, rather than the image as the other do. Maile66 (talk) 19:24, 6 February 2012 (UTC)
Hi, just started having an issue with the infobox and the centered image in Fluorine conflicting. It was NOT an issue for weeks before, just started right now. Seems to be only in IE9 an issue. Mozilla is fine and so is Chrome. And even IE was fine until just now (fine for several days).
FYI, moving the image to the left did not help (with space separators). Still have the conflict (basically image will not display until after the infobox is done). If I allow to text wrap, will work, but then I break the section header. REally I want it the way it was before (centered, looked sweet).
P.s. Don't fuss at me for too many pictures or a big infobox or centering an image. I like using the layout as more than just a wall of text and only text wrapped images. Help please instead of fussing. ;-) Really everything was fine until a few hours ago.
That was it. Turned it back off. Must have hit it by mistake...it is right next to the reload button. Thanks for your expertise!TCO (talk) 14:51, 7 February 2012 (UTC)
Hidden text
The edit box used to have a one-click feature to hide text. But, it's not there anymore. I forgot how to hide text, and I had to search for in a labyrinth of help pages. Why can't that feature be retained in the new edit box? Aditya(talk • contribs)13:26, 7 February 2012 (UTC)
"Hide text" can refer to different features and I'm not sure which one you mean. Below "Save page" you can select "Wiki markup" in a drop down box (if you have JavaScript enabled). There you can click <!-- --> to make a source comment if that is what you mean. PrimeHunter (talk) 13:42, 7 February 2012 (UTC)
Presumably a regex match is easiest, and to keep things simple (since I'm fairly confident we don't allow you to imitate anything that even looks like an IP address), we can use a more general regex that we might otherwise. If username is the relevant string:
If a set of four numeric values is tested, it's worth testing that each of those is in the range 0-255. That is, a user name like 256.256.256.256 cannot be an IP address. --Redrose64 (talk) 16:08, 7 February 2012 (UTC)
I don't think we allow people to register as User:256.256.256.256, though I could be wrong. and mw.user.anonymous would be fine if you have a user object; I'd forgotten about that. It depends how careful you want to be (AFAIK, it's all done by username anyway: there's no official IP/username divide). - Jarry1250[Deliberationneeded]16:28, 7 February 2012 (UTC)
Forgive me for not being very smart with this, but isn't mw.user.anonymous just returning whether or not I am an IP or not? What I need to know is whether a name I have stored in a variable is an IP. SpinningSpark16:50, 7 February 2012 (UTC)
How about mw.user.name() then (listed in the same manual page)? It apparently returns NULL for logged-out users. Ucucha (talk) 16:52, 7 February 2012 (UTC)
IP::isIPv6 is in turn defined at IP.php with some complicated regexes.
User::isValidUserName rejects usernames that match User::isIP, are blank or too long, contain a slash, or (for ucfirst wikis) begin with a lowercase letter.
I have tried to implement that regex at User:Spinningspark/monobook.js but I don't really understand regex very well. It seems to be returning userIP=true no matter what the input. Can someone take a look? - it is right near the beginning of the page. SpinningSpark19:06, 7 February 2012 (UTC)
Don't wrap the regex in single quotes. In other words, change '/^\d{1,3}\.\d{1,3}\.\d{1,3}\.(?:xxx|\d{1,3})$/' to /^\d{1,3}\.\d{1,3}\.\d{1,3}\.(?:xxx|\d{1,3})$/. — Bility (talk) 19:21, 7 February 2012 (UTC)
A common issue I've found is, someone will post a link to a talk page (or a village pump discussion or whichever), and then you click on it and get the most current version of the talk page with the content of that link archived. From that point you could try searching for that page, but finding an archived copy of a talk page based on a broken link can be painful. My proposal - I'm not sure. I mainly just wanted to say this is an issue in Wikipedia.
One possible solution, that would need fleshing out... Maybe next to section headers on talk pages/village pump pages there could be a link that says "permalink"? Jztinfinity (talk) 18:09, 28 January 2012 (UTC)
I've been moaning and groaning for years about broken/stale links to past discussions. I usually just end up manually fixing the links myself. -- Ϫ14:38, 29 January 2012 (UTC)
So ClueBot III changes links to archived discussion sections, but what about pages on which ClueBot III does not run, or pages in userspace which users do not wish to be changed? And what about external links (or even links kept personally by users outside of Wikipedia, say, copy/pasted somewhere)? The solution is not to change references to dead section links to point to different targets; one solution would be to design the software such that links to sections of discussion pages automatically redirect to archived copies. —danhash (talk) 14:55, 2 February 2012 (UTC)
Any link fixing bot cannot solve the problem of link rot because there are also links in edit summaries, such as in [1]. Let us wait for LiquidTheads to come? Incnis Mrsi (talk) 19:17, 8 February 2012 (UTC)
Deploying Swift for thumbnails this week (Feb6-9)
This week (Monday through Thursday), I will be switching over the backend system that hosts thumbnails (scaled images) for all wikis from our existing server to Swift, a clustered object store. This move gets us ready to be able to dramatically increase the amount of data we can hold in Commons and the other wiki projects.
As with all new systems, though it has been tested, the possibility exists that something will go wrong. I would like your help testing and reporting any issues. There are two main methods available to report issues:
Today, only files that contain "/a/a2/" in the URL are be affected. More files will be affected throughout the week following a gradual rollout schedule. Though I will take bug reports on other issues, they're less likely to be related to the change I'm making.
Ah, sorry 'bout that. Is there an article where you know for sure the edit notice was used? If so, and if the edit notice has been deleted, then perhaps an administrator would be willing to provide you with the deleted content so you could find your color. --Andrew(User:90)(talk)04:43, 8 February 2012 (UTC)
This search contains a lot of editnotices for TV episodes, but the couple I looked at did not have anything helpful in the history. I'm gonna have to give up on this one, but good luck finding what you're looking for. --Andrew(User:90)(talk)05:00, 8 February 2012 (UTC)
Automatically combining male and females into a single category
Is it possible to redirect categories? So if you put Category:Swordswomen on an article it would really go into Category:Swordsmen, but show as Swordswomen on the article? Richard-of-Earth (talk) 06:43, 7 February 2012 (UTC)
It's not possible to display the pages in the target category. See Help:Category#Moving and redirecting category pages. If Category:Swordswomen was redirected to Category:Swordsmen then an article using Category:Swordswomen would display Swordswomen on the article and clicking it would lead to Category:Swordsmen, but the articles using Category:Swordswomen would not be displayed there. PrimeHunter (talk) 12:44, 7 February 2012 (UTC)
The only permitted method for redirecting a category is by using the {{Category redirect}} template, see for example Category:Living People. If an article were given the wikicode [[Category:Living People]] (capital "P"), it would be categorised as such; and within 24 hours a bot will come along and recategorise it as [[Category:Living people]] (small "p"). I don't know if the bots will fixup all redirected categories though: Matekoraha Te Peehi Jaram has been in Category:Ngati Maru (Taranaki) for over three days now. --Redrose64 (talk) 14:06, 7 February 2012 (UTC)
This is a poor idea. It will just cause confusion if a page is displayed in a category but another category name is displayed on the page. PrimeHunter (talk) 20:40, 7 February 2012 (UTC)
Counting number of sections on page for transposing
I put something up earlier about this. I'm trying to get something that will count the number of sections in a particular page, and return the value to another page to be used with parser functions (so to return the nth last section.) I had a response that didn't help me earlier, so I'm bumping it back up. PuppyOnTheRadio (talk) 07:45, 7 February 2012 (UTC)
You don't need to start a new thread. If you post to your original thread, which is at #Counting number of sections on page, it saves the discussion from becoming split. This will become a greater problem when one of the two becomes old enough for archiving, but the other isn't. --Redrose64 (talk) 14:10, 7 February 2012 (UTC)
Good point. Still not an answer that helps me, and as the last discussion seems to have come to a schreeching halt, I'm just putting it up again now rather than waiting for the last one to be archived and then putting it up. PuppyOnTheRadio (talk) 00:56, 9 February 2012 (UTC).
The best I can come up with is:
Goto Google
If you don't have one already, open an account
Go to Google docs
Create a spraedsheet
Put in a cell this: =counta(ImportHtml(ʺhttp://en.wikipedia.org/wiki/Wikipedia:Village_pump_(technical)ʺ,ʺlistʺ,3))
You now have the count of topics from this page.
Put in a cell this: =ImportHtml(ʺhttp://en.wikipedia.org/wiki/Wikipedia:Village_pump_(technical)ʺ,ʺlistʺ,3)
You now have a list of topics from this page.
I don't know if this works for every article. You may have to change the index (3 in this case) for different pages. I leave it to you to figure out how to reverse the list. Richard-of-Earth (talk) 08:39, 9 February 2012 (UTC)
Template:US year nav is included at the bottom of every XXXX in the United States article. But it really makes a mess of itself in some of these articles, such as 1777 in the United States. Notice how the bottom elements of the page get all bunched up? I thought I would be clever and insert {{clear}} at the top of the template, purge everything, and then the problem would be solved. But that did not work. Does anyone know how this problem can be solved? Am I missing an obvious problem in the article page? --Andrew (User:90) (talk) 21:26, 7 February 2012 (UTC)
Ah, I see what you did. I just did the same to 1776 in the United States and it worked. I wonder if it is a good idea to use {{Div col}} at all for lists that only contain a few items. Perhaps this is not the correct place to ask, but would anyone object if I just removed {{Div col}} instead of closing it in the other XXXX in the United States articles that are unclosed? --Andrew (User:90) (talk) 21:46, 7 February 2012 (UTC)
Edit-counter pie chart for self use: I have been able to use the WikiChecker tool, but only for my own username and only with the default 500-edit analysis. For a user named "Axx" checking self:
When I tried to request an analysis of "1000" edits, it gave an error message, but at least it had given the pie chart for my recent 500 enwiki edits. -Wikid7714:55, 9 February 2012 (UTC)
{{Infobox ancient site}} has parameters for latitude, longitude and coordinates. Don't use latitude or longitude; use coordinates and the {{coord}} template with desired display options.
In Byllis, the problem with using |coordinates= instead of |latitude=|longitude= is that the pushpin map only works if |latitude=|longitude= are given. {{Infobox ancient site}} will display coordinates upper right when |latitude=|longitude= are given, provided that |coordinates_display=inline,title is also given.
Background: I'd like to be able to create wikitables that are sortable and present the athlete's age at a certain point, for instance at the beginning of a tournament. The Dts is sortable, but is not able to calculate and present age. The Bda2 calculates age and presents birth date and age, but is not sortable (well, it's sortable, but you know what I mean).
In other words, I'm looking for a template that essentially achieves this:
{{sort|1991-01-02|2 January 1991 (age 21)}} or {{sort|1991-01-02|January 2, 1991 (age 21)}} (depending on date format) from parameters date-of-birth, date-starting-point and date-format.
I don't know of any way to get it to calculate and sort at the same time. You'll probably have to go with the code you used inside the nowiki tags. Nyttend (talk) 03:28, 4 February 2012 (UTC)
I've created a shoutbox template that doesn't completely screw up your talk page, at Template:Shoutbox sidebar, with very simple and clean installation. People have been trying to use shoutboxes here since at least 2009, but they've always been weird floating boxes that got in the way of other content and features. Fixed! — SMcCandlishTalk⇒ ʕ(Õلō)ˀContribs.19:05, 3 February 2012 (UTC)
I added it at my talk. Can you make the "Click here" link edit "section 0", so that the rest of the talk page doesn't load when adding a message? —danhash (talk) 14:46, 10 February 2012 (UTC)
Transparent PNGs used for math formulas unreadable on black background
Math formulas created using Wikipedia's markup look fine on a white or light-colored background, but are virtually unreadable on browsers which are using a black background. This is because they are turned in to PNGs with black fonts (with a very faint gray outline) on a transparent background.
The only solution I can think of to this that accomodates users who browse Wikipedia with a browser configured to use a black background is to make all math PNGs use a white or light-colored background by default.
First of all, your lookpic.com links are HTTP 403. Please, point to a page in Wikipedia where such formulas on the black background may have some use. Even if it exists, a workaround for a particular task may be found easier than to accommodate some changes to texvc. Incnis Mrsi (talk) 13:26, 9 February 2012 (UTC)
403 for me as well. Some readers prefer a black background, such as can be enabled with Preferences → Gadgets → Use a black background with green text on the Monobook skin. If we switch to Monobook and enable that gadget, then we would probably see the issues. ---— Gadget850 (Ed)talk13:39, 9 February 2012 (UTC)
That gadget includes the appropriate workaround: img.tex{background:white;}. You can preview the gadget by adding useskin=monobook&withCSS=MediaWiki:Gadget-Blackskin.css to the url, e.g. [2]. Anomie⚔13:58, 9 February 2012 (UTC)
I might not call this workaround an appropriate one. White bars with black glyphs are definitely better than black glyphs on black background, but this does not yet fit to the green-on-the-black style. File a bug to bugzilla: to permit custom foreground in WP:texvc. Not only green and white foregrounds can be used, but possibly also a blue one, to place hyperlinks on formulas. Incnis Mrsi (talk) 15:12, 9 February 2012 (UTC)
How is the math renderer going to know what the background color happens to be, to pick the right foreground color? Or if you're talking about specifying the color in wikitext, that won't help much either as white-on-white for the vast majority of readers is worse than black-on-black for a few who know what they did. Unless perhaps you're proposing some mechanism where a user Javascript can swap out all the black-text images for white-text images? Except that the way the Math extension is structured, the image is rendered just once when the wikitext is parsed (unlike image thumbnails, which are rendered when the HTTP request for that sized thumbnail comes in). Anomie⚔20:41, 9 February 2012 (UTC)
Unfortunately, this solution does not really help me. I use my own stylesheet when looking at Wikipedia, and even were I to use Wikipedia's stylesheet, in order to configure it to use a black background I'd need to sign in to use it -- something I don't do 99% of the time. I only sign in to edit, which is pretty rare these days. As such, solutions which require me to sign in to use are pretty useless. That probably goes for most other users who view Wikipedia with a black background, as most of them probably don't even have Wikipedia accounts in the first place. -- noosphere17:04, 9 February 2012 (UTC)
What an excellent suggestion! It worked! Here is how a formula from the Mathematics article looks like now. Not perfect, as the top of the radical symbol and the tops and bottoms of some numbers look cut off due to them being too close to the edge of the white background (and I'm sure I'll run in to other similar border issues with other formulas), but it's far better than the alternative. This is also probably not the best solution for Wikipedia in general, because it will require users who haven't logged in but who use a black background in their browser to change their personal stylesheets, but at least now it's possible. Finally, of course, the remaining objection is one that's been raised before: that it looks ugly. It would look much better if the foreground color could be changed to white instead. However, I don't really mind, as I'm perfectly capable of changing my own stylesheet and as long as the results are reasonably readable, I am happy. -- noosphere08:58, 10 February 2012 (UTC)
Sorry about the 403 error. Hopefully the following imageshack links will work: white bg, black bg. As for your request to see "a page in Wikipedia where such formulas on the black background may have some use," that would be any page with a math formula! For example, the formulas on the Wikipedia article on Mathematics look pretty much like the ones I linked to above: virtually unreadable to someone viewing the article with a browser configured to use a black background. -- noosphere16:58, 9 February 2012 (UTC)
Which browsers are these that override the Wikipedia house style? The only browser that I've used where black background is the default was Lynx, a text-only browser, where image colours are a non-issue. --Redrose64 (talk) 21:43, 9 February 2012 (UTC)
It's not the default, but both Firefox and Opera allow you to override any site's stylesheet. For example, Firefox has an extension called Stylish (with close to a million users), which lets you choose from a variety of global and site-specific styles -- some of the most popular are styles that use a dark background for browsing at night. Right now I'm using another Firefox extension that allows me to override default styles. The Opera browser has a number of built-in styles that could be selected to override the any site's default style -- some of these styles have black backgrounds. I'm sure there are other ways of overriding stylesheets in Firefox, Opera, and most other browsers. -- noosphere09:20, 10 February 2012 (UTC)
I added the block code from Template:Wikipedia ads to my common.css, and it worked. Then it randomly stopped working, so I added !important, and it worked again. I am now again seeing these stupid ads on talk pages, and I want them gone. Forever. Do I even have to explain how annoying and pointless they are? Anyway, what must I add to my common.css to never, ever, ever, ever see these again? (common.css only works while logged in of course, so I'll also add it to my Adblock Plus.) —danhash (talk) 18:22, 9 February 2012 (UTC)
MediaWiki either moves the wrong line outside the table or removes it. In this case the wrong line was another table definition {| class="plainlinks qxz-ads" so after it was removed the CSS class was simply not there. — AlexSm17:00, 10 February 2012 (UTC)
I'm trying to download John Williams as a PDF (using this link from the toolbox (Download as PDF)). However, when I download it, it appears as unformatted plain text with the error message "WARNING: Article could not be rendered - outputting plain text. Potential causes of the problem are (a) a bug in the pdf-writer software (b) problematic Mediawiki markup (c) table is too side."
The precise text in the generated PDF file (including a typo in "ouputting" which is needed to search other reports): "WARNING: Article could not be rendered - ouputting plain text. Potential causes of the problem are: (a) a bug in the pdf-writer software (b) problematic Mediawiki markup (c) table is too wide".
It's the table that's causing the problem. I made a copy of the article in a sandbox page and commented out the table and it rendered fine. I'm not convinced it's the table's width that is the problem though (I tried a different page with a full-width table, and that worked fine); my guess is that it's the table's overall size causing the issue. But having said all that, I don't know a solution either. Sorry. DH85868993 (talk) 01:51, 10 February 2012 (UTC)
With respect to commons we usually don't keep local copies of free files available globally and have on the other hand some fair-use files they won't have. What about a third situation where a file is available on commons but it might not fit in here for any reason such as being a POV version of another file, blatantly promotional etc. Leaving apart for a moment the question of criteria, let's simple assume there is a file that we would delete here, but commons not, so it exits there. Are there any technical means to prevent its inclusion here? --Tikiwont (talk) 21:43, 10 February 2012 (UTC)
Thanks. The first is akin to the spam list that i had in mind and the second might be a useful occasional quick fix. --Tikiwont (talk) 22:08, 10 February 2012 (UTC)
Interwiki bots
Interwiki bots stubbornly re-add interlang links that humans have judged are inappropriate. This is unacceptable. All interwiki bots should be reprogrammed to never re-add an interlang link that a human editor has removed. --Trovatore (talk) 22:41, 10 February 2012 (UTC)
Hmm, drastic perhaps, but how do we publicize the HTML-comment thing, and does it always work? I completely disagree with the notion that editors should have to go around to all other languages and fix them so that the interwiki bots don't do stupid things. That should be the responsibility of the bot writers and operators, and if they fail in that then there needs to be a way to turn them off. --Trovatore (talk) 04:15, 11 February 2012 (UTC)
I'm not sure why the section edit tabs are not showing for sections 9 (Articles 101 through 120) through 19 (Articles 301 through 320) of Wikipedia:Contributor copyright investigations/PumpkinSky. I'm currently doing some work in this area, and editing is quite slow because I have to save such a large section. I suspect it is something quite simple, but I'm not seeing it.--SPhilbrick(Talk)18:37, 11 February 2012 (UTC)
Fixed There were two sets of stray double braces that interacted with {{cot}}/{{cob}}. When you would edit section 9, the following sectioos were in the edit window as if they were subsections. ---— Gadget850 (Ed)talk19:30, 11 February 2012 (UTC)
There's something wrong with whatever MediaWiki page governs the format of the links to create editnotices in the top right corners of editing windows: while the "Page notice" line is working fine, something's suddenly gone wrong with the "Group notice" link — instead of displaying simply as "Group notice", it now displays as [[Template:Editnotices/Group/Template:FULLROOTPAGENAME|Group notice]]. Any idea what's wrong? I don't get any relevant edits from Special:RecentChangesLinked, and (except for one page that was edited nine months ago) all of the MediaWiki pages that link to Template:FULLROOTPAGENAME were last edited in 2009. Nyttend (talk) 02:41, 12 February 2012 (UTC)
Apologies if this has been covered elsewhere. When patrolling the unpatrolled backlog, I've found a few pages which've been moved without a redirect which are still appearing on the backlog. These include Swain (Ski Area) and Apcpdcl. As the pages don't exist, there's no way to patrol them and so they seem destined to recur at the bottom of the backlog for eternity. I'm aware this is hardly a major issue—I was mostly just wondering if there's an obvious way of marking such pages as patrolled that I'd missed &/or if there's a workaround of some sort already in place? Thanks. – hysteria18 (talk) 23:39, 12 February 2012 (UTC)
You can purge the page by appending &action=purge to the New Page log url (for example, here). This will remove the deleted pages from the log. Goodvac (talk) 23:57, 12 February 2012 (UTC)
When I use Windows Media Player I get : Windows Media Player cannot play the file. The Player might not support the file type or might not support the codec that was used to compress the file.
When I use Quicktime I get : Error -2048 Couldn't open the file Berceuse_by_Gabriel_Fauré_op56_no1.ogg because it is not a file that Quicktime understands
I've downloaded the Codex thingy and associated .OGG with Windows Media Player and can now successfully play the downloaded .OGG file; however when I just click on the button from the article I get a little red cross... GrahamHardy (talk) 03:12, 13 February 2012 (UTC)
Obtaining a recent changes feed
I have started working on an anti-vandalism tool that I hope will replace Lupin's tool (which was created long before the API existed), and I am questioning whether RSS (or Atom) is still the best way to obtain recent changes information, including diffs. There is no way to specify the starting time (or better, rcid) of the feed, meaning either edits will get lost or will be sent twice or multiple times. On the other hand, I could poll the API for a list of changes (or use the IRC feed if possible) and then download the diffs individually. I would like some feedback on which is the better approach and if there is an approach I have not yet thought of. PleaseStand (talk) 13:42, 13 February 2012 (UTC)
The message Mediawiki:Addedwatchtext should be corrected, it doesn't show an appropriate text if one adds a template talk page to his or her watchlist. For example, when I'd added the Template talk:Space on my watchlist, the message: "The page "Template talk:Space" has been added to your watchlist, which will list edits to this page and its associated talk page." appeared - that doesn't really make much sense. --Eleassarmy talk17:10, 13 February 2012 (UTC)
It is possible to use something like " associated {{#ifeq: {{NAMESPACE}} | {{TALKSPACE}} | subject | talk }} page". — AlexSm17:23, 13 February 2012 (UTC)
Hi, the presence of the sound clip on today's main page once again highlights the unsatisfactory nature of Wikipedia's media handling. Every single website in the universe, with the sole exception of Wikipedia, plays media without any problems on my PC. Only Wikipedia has some bizarre and arcane method that requires installation of God knows what unknown software from some source I have never heard of and do not by default trust. It is completely crazy. I believe I read an explanation once, but whatever that explanation is, it needs serious reconsideration. 86.177.105.189 (talk) 18:48, 13 February 2012 (UTC)
I've downloaded the Codex thingy and associated .OGG with Windows Media Player and can successfully play the downloaded .OGG file; however when I just click on the button from the article I get a little red cross... GrahamHardy (talk) 03:12, 13 February 2012 (UTC)
The individual Toolserver web servers (wolfsbane and ortelius) do have IPv6 addresses, which are 2620:0:862:101::2:6 and 2620:0:862:101::2:7 according to DNS. PleaseStand (talk) 12:59, 14 February 2012 (UTC)
The message I got today was:
The RIPE Database is subject to Terms and Conditions.
Apparently, if the server is registered as a proxy (which the toolserver tool is functioning as in this case), you can do 20,000 queries per day rather than just 1000. [3]Reaper Eternal (talk) 20:27, 14 February 2012 (UTC)
Could someone with dev access delete this page please? WP crashed while I was editing it, and now it hangs every time I try to bring the page up. - Dank (push to talk) 19:29, 14 February 2012 (UTC)
You could have deleted it yourself, just append "?action=delete" to the standard URL to bring up the deletion form without having to load the page. Anomie⚔19:58, 14 February 2012 (UTC)
So I've currently got a javascript/jquery code (some of which I understand, most is gobbledygook) that I use to call and manipulate the top 10 entries in my watchlist, to display on a page where I keep all my daily stuff to check. It got me thinking, would it not be possible to modify this code to also query my commons watchlist, then combine the two arrays/tables and list by order of timestamp?
Unfortunately I suck at javascript, and even when I try to make a query to commons nothing happens.
Drawing images with a markup language - is it possible now or in near future?
Hello! In Wikipedia you can write equations in LaTeX:
but is it possible to draw images with a markup language? And if it is not, do you know if such a feature will be implemented in near future? It would be a significant improvement for:
image coherency (for example, if you look the electrical engineering articles in Wikipedia, all the circuit diagrams have different appearance because they are drawn with many different tools. Using one markup language would harmonize the outlook of the diagrams and make Wikipedia look more professional. Same applies to block diagrams etc.)
ease of image editing (It is easy to fix little mistakes (text in images etc.) or update the images by just editing source code, instead of fighting with image editor.)
First of all, using one markup language would not harmonize the outlook of the diagrams, and also could make Wikipedia look inconvenient for editing and dull, as there are deliberately much of diversity here. Don't invent anything new, while there is something old to apply, for you. Incnis Mrsi (talk) 09:14, 14 February 2012 (UTC)
I disagree. Adding such a markup language probably would help standardize the look of electrical diagrams, for the same reason regular TeX standardizes the look of Mathematical equation. Plain text or other format equations are acceptable, but they may be converted. People would not be required to use the language, any more than they're required to use TeX or SVG. I don't agree that providing a useful tool makes Wikipedia "dull". Superm401 - Talk16:54, 14 February 2012 (UTC)
I know SVG. As a TeX package, TikZ provides a different approach, which some users may prefer for mathematical or scientific diagrams. For instance, it can allow GNUPLOT integration. Superm401 - Talk06:39, 15 February 2012 (UTC)
"regular TeX standardizes the look of mathematical equation"? It is true for mathematical and physical journals, but applied to Wikipedia… is sounds like a fairy tale for children's bedtime. I know that coders cannot fix a bug with HTML "−" presentation for seven years, and I made numerous edits to replace <math> with explicit HTML to get rid of texvc's ugly (HTML rendering must be enabled to see what is wrong here). And I saw numerous instances of bad typesetting despite the use of <math>. Incnis Mrsi (talk) 17:26, 15 February 2012 (UTC)
I am not saying Wikipedia's TeX implementation is perfect. I support Vesa's idea to use the TikZ package for additional diagram support. But I also support bug fixes to the existing system. Superm401 - Talk23:09, 15 February 2012 (UTC)
Thumbnail wrongly generated
Help! Last week I updated a file I had uploaded for use in Template:Contains_Korean_text, which appears in dozens of articles. I checked that my change was appearing at the time and all was well. Today I find that my correction appears to have reverted in the infobox! But, going to the file’s page, the full size graphic displays correctly; only reduced versions are wrong. (The 120px thumbnail listed under versions on the file page, and the 40px version in the infobox itself. The difference is the character at the lower-right.) Is there a way to force thumbnails to regenerate? And how could this have happened? MJ (t • c) 06:35, 15 February 2012 (UTC)
I recently created an alt account at the suggestion of an admin. Which is User:Edinburgh Wanderer Mobile I'm trying to find out if there is a way my watchlist on this my parent account can be transferred over to the mobile account without having to do it manually. EdinburghWanderer10:36, 15 February 2012 (UTC)
There is a link to the editable list on your watchlist page: Edit raw watchlist. You can copy this list of watched titles, then switch accounts, and paste the text in to your new raw watchlist list. — Richardguk (talk) 11:27, 15 February 2012 (UTC)
Is it possible to block an editor just from some pages but not others?
I mean with MediaWiki in general, not necessarily the way it's configured on en.wp. I'm asking this question because of this ArbCom proposal, in particular the community is encouraged to review and document common good practice for administrators imposing editing restrictions as a condition of an unblock and in lieu of blocks.ASCIIn2Bme (talk) 20:07, 15 February 2012 (UTC)
I know that .js modifications to the UI will load on the watchlist, so that is not entirely correct. I also know that they will not load on the preferences page. Presumably to stop people from breaking their own UI. :D --Izno (talk) 23:43, 15 February 2012 (UTC)
As far as I know Special:Preferences has always disabled user scripts for security reasons. A script could change your email and then the script-writer could request your password sent by email and take over your account. — Bility (talk) 00:44, 16 February 2012 (UTC)
Could the job queue be stuck? It's been a long time since I looked at this page, so I don't remember what normal is. I ask, because things slowed about 22:33 on the clock. That's how long I've been waiting for changes to show on "What Links Here" on a couple articles. Things were going fairly rapidly before that. Maile66 (talk) 23:50, 15 February 2012 (UTC)
Hmmmmmmmmmmm. And 25 minutes since I posted this, it still doesn't show up on my Watchlist as anything changed on the Pump. Maybe it's a busy server. My server seems to be "mw 45". Maile66 (talk) 00:17, 16 February 2012 (UTC)
OK, everything finally cleared/showed up on my end like it's supposed to. That's an 2 hours and 45 minutes for the Watchlist items to show up. Whatever...it's OK now. Maile66 (talk) 00:22, 16 February 2012 (UTC)
Instead of disabling it, can I change it so that it links to my actual sandbox rather than a non-existent page? Roger (talk) 12:18, 16 February 2012 (UTC)
mw.util.addPortletLink('p-personal','/w/index.php?title=Special:MyPage/Sandbox&action=edit&preload=Template%3AUser_sandbox%2Fpreload&editintro=Template%3AUser_sandbox','My sandbox','pt-mysandbox','Go to my sandbox');
The preload adds {{User sandbox}} if the sandbox does not already exist. The editintro shows the same message. If you don't want them, then use:
mw.util.addPortletLink('p-personal','/w/index.php?title=Special:MyPage/Sandbox&action=edit','My sandbox','pt-mysandbox','Go to my sandbox');
If you don't want it to open for editing, then remove &action=edit
Does anybody know how to programme your sandbox monobook to place a border frame, sort of like a gold picture frame around the edge of the wiki screen? I just want to embolden it from my browsing panels that's all.♦ Dr. Blofeld18:11, 16 February 2012 (UTC)
Are you looking for a CSS implementation, or would you be willing to wrap your whole sandbox in a <div>...</div>? What I mean is, do you want others to be able to see the border also? — Bility (talk) 18:32, 16 February 2012 (UTC)
My bad, I meant monobook LOL. In my preference I want to be able to have a frame around the pages or the whole thing to embolden it in contrast that's all.♦ Dr. Blofeld18:51, 16 February 2012 (UTC)
Not sure if I'm in the right place, but after using {{help me}} I was advised to ask my question here.
When logging into Wikipedia through unified login, I'm logged into every project except Wiktionary(see revision in reply of February 10) (and of course the three two projects pending SUL-request). At first this was solved by not using the secure server (https://), but now it refuses regardless of settings. When logging into Wiktionary directly there is no problem, but then I won't be logged into any of the other projects. In both cases the icons shown suggest I should be logged into all projects. To make it even more interesting: when logging into Wiktionary (manually) after logging into Wikipedia, then logging out on Wiktionary, all of a sudden I am logged out on Wikipedia too, so for logging out it does seem to work... I haven't been able to find anything in the help sections, does anybody have a suggestion? Thank you, — Quibus (talk) 14:00, 8 February 2012 (UTC)
Curious. When I'm logged into Wikipedia, going for wikt: shows the "Redrose64 / My talk / My preferences / My watchlist / My new messages (None) / My contributions / Log out" links upper right, so I'm logged in there too. If I then log out of Wiktionary, and return to Wikipedia, I find that I have been logged out there too. This is what I expect SUL to do.
I have observed the following over a period of many months, which may not be the whole story. The process of logging out zaps the cookie created when you logged in. I think that SUL means that one cookie is shared by all projects: therefore, logging out of Wiktionary will also log you out of Wikipedia. When additional projects join SUL, the user database is not always copied over, so it's worth visiting Special:MergeAccount every so often to keep them up to date.
When the secure server had its own URL structure, i.e. https://secure.wikimedia.org/wikipedia/en/wiki/Main_Page for the Main Page, it had its own cookies. Now that the only difference in URL structure is https versus http (the secure server main page is now at https://en.wikipedia.org/wiki/Main_Page), it can be demonstrated that by logging in "insecure" and then switching to "secure" (by altering http to https in the browser address bar) you are still logged in, therefore the cookie is shared. --Redrose64 (talk) 15:38, 8 February 2012 (UTC)
The cookies are only shared if you log in insecurely. If you log in using https, the cookies are marked "secure" so you will not be logged in when accessing via http. Anomie⚔17:30, 8 February 2012 (UTC)
Thank you both for your input and suggestions. I have visited Special:MergeAccount every so often since my SUL-request was initiated. After reading your posts I've checked whether removing cookies would do the trick (it didn't) and if logging in through secure server again would change anything (it also didn't). When checking the projects in the Special:MergeAccount list on whether I'm logged in or not I noticed Wikitionary is not the only project I'm not logged in, it's every project under Wikimedia I'm not logged into and a couple of foreign Wikis:
Unified login succesful: af, en, es, fo, fr, ga*, gd*,it, nl, pl, ru, tpi, uk
You have accounts at gawiktionary and gdwiktionary but not gawiki and gdwiki, so ga and gd in your Unsuccesful must be at wiktionary. Then it appears Wikipedia is treated specially for you. Some browsers can make cookie settings for individual domains. Maybe you accidentally did this for wikipedia.org. Which browser do you have? Can you try another? PrimeHunter (talk) 12:57, 10 February 2012 (UTC)
Thank you PrimeHunter, you're absolutely right about the ga and gd account being Wiktionary, I didn't notice. After visiting them again (still not logged in) I also visited ga-wiki and gd-wiki, where the unified login did work, so they both appear in the Special:MergeAccount list now.
I didn't think of using another browser, thanks for that one, sadly it doesn't make a difference. Standard browser is Firefox 9, alternative is IE9. I'll try checking cooky settings now, in the mean time any suggestion is welcome. — Quibus (talk) 16:16, 10 February 2012 (UTC)
Resolved! Looking at the cooky settings I noticed I had third party cookies disabled. Enabling it did the trick and everything works fine on https:// too. Thanks again everybody for all the input. — Quibus (talk) 19:09, 10 February 2012 (UTC)
I've found that Special:MergeAccount doesn't always pull in all other projects. If you check http://toolserver.org/~luxo/contributions/contributions.php?user=Xxxx (where Xxxx should be replaced by your username) you may sometimes see that one or more of the listed projects has a pink background; near the top it says "SUL: Account unattached" instead of "SUL: Account automatically created ...". Click on any page listed in the pink region, go for your browser's "back" button then the "refresh" button, and you should find that the pink background disappears indicating that your accounts are now linked. --Redrose64 (talk) 14:23, 17 February 2012 (UTC)
This is unrelated to Quibus' problem. What you describe is almost certainly cases where a page you have edited has been transwikied to a wiki where your account is not registered yet. Just like other wikis, your account is automatically registered first time you visit any page there. I admit it's confusing that your edits can be registered in a wiki where your account is not registered. PrimeHunter (talk) 21:37, 18 February 2012 (UTC)
I have to report that now that I've unticked MWEmbed the padlock is green again in Chrome (although I could have sworn that it continued to be red initially after removing the gadget from my preferences). It Is Me Heret / c13:59, 17 February 2012 (UTC)
The server which is hosting that gadget, unfortunately isn't available on https atm... It's in the works of going trough code review, so i doubt it will be updated before that is finished... I'll ask anyway though. —TheDJ (talk • contribs) 12:09, 18 February 2012 (UTC)
Printable version without images
Is there a way to generate a printable version of an article, that would not include images (and their descriptions)? --eugrus (talk) 19:19, 13 February 2012 (UTC)
This will suppress images. Need to dig a bit to get the captions. Will get back to this in a bit.
/* do not print images or captions*/@mediaprint{img{display:none;}.thumb{display:none;}.thumbinner{display:none;}.thumbcaption{display:none;}.magnify{display:none;}}
/* do not print images or captions*/@mediaprint{img,.thumb,.thumbinner,.thumbcaption,.magnify{display:none;}}
Works also and requires less css. Not that it matters because it ends up being minified (to some extent; mostly the whitespace), but just noting. --Izno (talk) 19:53, 14 February 2012 (UTC)
I am now getting where the misunderstanding is: what I am talking about is getting rid of images on "&printable=yes" pages: not just in the actually printed text on paper. Is this possible? --eugrus (talk) 22:20, 15 February 2012 (UTC)
How does one go about getting a new filetype added to the list of allowable uploads?
The title pretty much says it. I don't want to spill the details as we are still working on a model, but its an XML based filetype (aka text based) that could provide an untold benefit to the project. Who deals with this? I know there will need to be community consensus to implement it, but when that has been achieved where does it go? Bugzilla? - ʄɭoʏɗiaɲτ¢17:33, 15 February 2012 (UTC)
Yes, Bugzilla. Do you mean creating an entirely new filetype (as opposed to adding an existing filetype to the list of allowed filetypes)? Ucucha (talk) 21:04, 15 February 2012 (UTC)
Thanks Alex; knew I'd seen it proposed somewhere recently. Are interface changes such as this one documented anywhere that's easy to find? When searching for documentation about the change I came up empty. —danhash (talk) 20:59, 15 February 2012 (UTC)
As far as I know, there is no formalized way in which interface changes as made by the English Wikipedia community are published centrally, though usually, this very spot is a good place to keep an eye on. —TheDJ (talk • contribs) 12:21, 18 February 2012 (UTC)
As a result of an RfC at WikiProject:Highways, a discovery was made recently that allows us to display linear data on Google Maps and Bing Maps with an extraordinary level of detail. Thanks to User:Dschwen, WikiMiniAtlas already supports the new data. In addition to the simple linear features that we have exploited thus far, there exists the potential to label specific points, or display entire networks (think subway systems, or rivers and their tributaries) on the mapping services. Lots of potential, but still lots of groundwork to lay.
To see a few examples of what this can do, check out:
River Liffey (link in title), a convoluted river in Ireland
Deansgate (link in the Further reading section), a short urban major roadway that begins and ends inexplicably
Note how in River Liffey we can put markers on the various landmarks along the river's course, as well as including separate lines for the tributaries that feed the river.
The main issue faced thus far is manipulating the data, as we cannot upload a *.kml file. Since they are text-based XML files, the contents can be copy-pasted into a subpage; however this presents a significant hurdle to user-friendliness.
Google Earth, NASA WorldWind, Gaia, and most GIS software can open, manipulate, create and save these files, so they are far easier to create than the standard PDF format. However, the files are article specific, so I feel they aren't appropriate for commons in most cases.
Support—as a supporter of the KML-based solution, I feel that we need to take this additional step before pursuing the concept on a wider range and number of articles. As it stands, we have no way to easily document the source of the KML data, and using subpages makes it more difficult for readers to reuse the data files. Imzadi 1979→02:50, 17 February 2012 (UTC)
Given that they are subpages, I'd assume they were manually added by copy/pasting the contents of a kml file, similar to what we are doing here at the moment. - ʄɭoʏɗiaɲτ¢22:16, 17 February 2012 (UTC)
I take your point. From the instructions at Geocoding/Overlay, you are right. It hadn't occurred to me that the difference is significant, but I suppose it means that the existing KML "files" are in fact merely "file description subpages" containing KML code (as their wikitext) but which, as far as the servers are concerned, have no file attached. (Afterthought: In some ways, a file description page is more accessible than a proper upload, being editable in place and simple changes viewable as diffs. But I concede that it is a bit of a kludge.) — Richardguk (talk) 23:32, 17 February 2012 (UTC)
Support. The current way of using subpages is actually fine, too. But if you guys want to upload the data as files on commons, that would be fine, too. It would just be a bit more difficult for me to obtain the data (due to crosssite scripting prohibitions (unless accessin via Special:FilePath works...)). KMZ would require me to write a decompressing proxy on the toolserver. I would certainly prefer raw uncompressed XML/KML data, as that is very easy to process in JavaScript using XMLHTTPRequests :-). --Dschwen23:37, 17 February 2012 (UTC)
Actually, if upload.wikimedia.org serves the file with a Content-encoding: gzip header I wouldn't have to worry about unzipping it, the browser would do it for me. --Dschwen00:02, 18 February 2012 (UTC)
Graphic for portal link to Academy Award portal isn't loading
When the following is added to articles to link to the Academy Award portal, the graphic isn't loading, and the generic portal image appears:
Is it possible to add RSS feeds of external pages to Wikipedia? If so, how do I do it?
For example, I would like an updated feed of my favorite CDs from Listal.
I know the <rss> tag works on Intellipedia, but what will work here?
Thank you for your help. Allen (talk) 17:48, 18 February 2012 (UTC)
So, unfortunately, it looks as if it can't be done. Does anyone know if RSS feeds from Wikipedia pages can be put on Listal? Allen (talk) 21:48, 18 February 2012 (UTC)
Looking for a tool
For years now, I've been looking for a tool which would be immensely helpful in matching suspected sockpuppets to possible puppet masters. What I need is a tool into which you can put a number of articles, and it would return the names of editors who had worked on all those articles, or a significant subset of them -- this on the theory that socks frequently return to the "scene of the crime" and can't resist editing the articles they edited under previous IDs. The list generated would provide some starting points to look for behavioral commonalities.
Wikistalk (which is a great tool) actually does the opposite of what I want. It takes multiple editors and tells you what articles they have in common. What I need is a tool which you give articles and it generates users who edited them in common, Beyond My Ken (talk) 05:04, 18 February 2012 (UTC)
You'll want to filter out a lot of bots and other prolific editors, but otherwise, yeah, it sounds straightforward. Got any particular articles in mind? 67.117.145.9 (talk) 00:40, 19 February 2012 (UTC)
Quartic function has spam inserted directing all clicks to a racist 9/11 conspiracy site
At least for me, going to the page quartic function leads to a strange situation where all clicks on any link (including "edit", "talk", etc.) lead to a racist Jew-baiting 9/11-conspiracy site at pump.pp4l.me. Conceivably this is a virus on my machine but I suspect it's code inserted into the page itself. I can't edit the page or look at its history, but I can see the page source, and I see the following:
This is stuck in directly between "This suggests using a ..." and "... resolvent cubic whose roots may be variously described as a discrete Fourier transform or a Hadamard matrix transform ..."
(This is one of the things that suggests a page mod rather than a virus -- it occurs only on one page, and it makes use of Wikipedia-specific links, and it inserts in the middle of some text)
Note that the div places an invisible image over the entire window, intercepting all links. Please fix this and permanently ban whoever put this in!
That was a quite devious abuse of the template mechanism. The "link" mechanism used there looks quite dangerous: is there ever a valid use for it other than to link to other pages on Wikimedia projects? If not, perhaps it should be filtered in software at render time. -- The Anome (talk) 11:05, 18 February 2012 (UTC)
Filter 139 is currently disabled, apparently because of too many false positives -- otherwise, it looks likely it would have caught this. Perhaps we should have a variant of filter 139 that just checks edits in template space? -- The Anome (talk) 11:27, 18 February 2012 (UTC)
The really bad thing here was to be able to make images (in this case, an invisible image) link to external sites. A whitelist for the linking mechanism in question would have defeated its principal purpose of redirecting the user to an attack page. I wonder, is there ever any legitimate reason to use this feature to link anywhere but to the same wiki the page itself is on? If not, we could restrict the attribute to use only relative links. -- The Anome (talk) 11:44, 18 February 2012 (UTC)
(edit conflict)When I talked about my edit there, in 2010, the filter was changed to allow established users pass through. I have not followed this issue after that. Filter usage is sensitive wrt the workload, very often there is a tradeoff. -DePiep (talk) 11:47, 18 February 2012 (UTC)
Let's make Filter 139 specific to the template space, then, and re-enable it. Perhaps we should even make the exception admin-only, rather than autoconfirmed? However, I'd still like to see action on making the "link=" parameter safe, as that is a long-term and complete fix to the root problem of this exploit. -- The Anome (talk) 11:53, 18 February 2012 (UTC)
The link parameter is likely to be used in template space. So perhaps a filter that looks for the combination 'link=' and 'position:absolute' would be more effective. — Edokter (talk) — 12:16, 18 February 2012 (UTC)
That is also used in templates. One thing that is used in vandalism such as this is an image linking to an external site - does this have any legitimate uses? Peter E. James (talk) 21:08, 19 February 2012 (UTC)
Add position:fixed as was used in this current abuse. Is the third aspect, the option "transparent image covering whole page" prevented this way? -DePiep (talk) 12:41, 18 February 2012 (UTC)
Okay, they hit a new template with a different but similar link. Anything from the above discussion viable to prevent future attempts? Ravensfire (talk) 17:24, 18 February 2012 (UTC)
Note: We've been getting a lot of complaints about this at the Help Desk. See the following threds:
On a related note, we are still getting questions about this at the Help Desk as people need to clear their cache in order for the updated, non-vandalised pages to appear. Would it be possible to temporarily put something in place that either did this for people or told people to refresh their browsers to deal with this? I'm not incredibly well-informed, so I don't know exactly what would be needed, but something would be helpful. ItsZippy(talk • contributions)19:13, 18 February 2012 (UTC)
I've now adjusted filter 139 to catch at least some of these, and enabled it. Please could someone check my edits to the filter for sanity, and check that they catch the edits in question? -- The Anome (talk) 19:23, 18 February 2012 (UTC)
I have not checked the HD links.
imo, there are three orifices of attack:
OR1: Cover the whole page with an invisible, linked image
OR2: Link to an external page
OR3: Use style="position:fixed" or ..."position:absolute"
Then there are angles of prevention:
ANG1: Use a filter to prevent some code (like filter 139; could use user-status)
422 deals with this as well, stemming from a similar set of attacks last month, though apparently not quite comprehensively; I've emailed the author of that to take a look at it. Shimgray | talk | 19:44, 18 February 2012 (UTC)
I'm glad I found this thread (after stumbling upon a link at ANI). Was trying to figure out where to go to, to point out this weird phenomena. I was reading the London 2012 Olympics article and noticed this strange redirect to a 9/11 Jewish conspiracy page, whenever I clicked on a link, image, or even just the plain white background of the article. I'm highly concerned, as the site it directs me to is causing my anti-virus programme to go berserk, telling me that my computer is under threat. Has anyone else's anti-virus software been flashing up with warnings from this weird site? Wesley☀Mouse20:00, 18 February 2012 (UTC)
Thanks for adjusting the abuse filter to stop / slow this down. For those that find a page affected by this but aren't sure were it's from, edit the page and look at the templates used in that page. Look at the history for each template that's not protected and you should be able to figure out which one was changed fairly quickly and revert it. Ravensfire (talk) 20:16, 18 February 2012 (UTC)
I've just created and enabled filter 453, which should temporarily stop all recently-created editors from editing templates at all. I'll disable it in a few hours, when this has blown over, and all the affected templates tracked down and fixed, but I think it will be a good idea to keep it around for the time being so that it can be re-enabled at a moment's notice if this starts up again. -- The Anome (talk) 20:37, 18 February 2012 (UTC)
OK: this all seems to have gone quiet for now, and I can't see any more reports of similar vandalism recently. Filter 453 is now deactivated, and filter 139 will stay active for now to watch out for similar attacks. If this recurs, please let me and WP:AN/I know, and further action will be taken. -- The Anome (talk) 22:28, 18 February 2012 (UTC)
OK. I've now deactivated 139, and reactivated 453, which is now hopefully both simple and specific enough to work. The oversighting of the abusive diffs, and the difficulty of extracting diff content without getting an eyeful of hate/malware content, has made this very difficult to debug. -- The Anome (talk) 01:34, 19 February 2012 (UTC)
There is an option in the Misc tab of My Preferences for "Do not show page content below diffs", although I don't know if it works for deleted revisions. It also seems strange that the diffs would be oversighted (and not just deleted via WP:REVDEL), as some of the links are still visible via the spam blacklist. Peter E. James (talk) 21:08, 19 February 2012 (UTC)
I've created a template at Commons (link) that's meant to inform the reader about a lot of copyright information at once. Unfortunately, because copyright is very complicated, there is a ton of template logic inside. The result is that any page which uses the template takes about 7 seconds to load. Is there a way I might cut down on the load time of this template? I cannot think of one, other than forking the whole project into JavaScript and adding it to the site's vector.js. Magog the Ogre (talk) 19:31, 19 February 2012 (UTC)
Of course, you might want to hold off on the port to JS (if it's required in the end) until Lua support goes live (tentatively expected to be near the end of this year), although tangible case studies may induce faster work on the project). - Jarry1250[Deliberationneeded]13:32, 20 February 2012 (UTC)
Cannot login from iPhone
I have tried during today to use my iPhone's browser to log in and make some edits. This fails on multiple accounts, regardless of which browser I use (tried Safari, the Wikipedia app, and Grazing, a third-party browser; but I suppose they just differ in the chrome, not in the actual HTTP(s) client).
Firstly, there is no "log in" link in the mobile version of Wikipedia. This may be a conscious decision (even as to dissuade mobile users from attempting to log in because of support issues, perhaps??) but I am mentioning this in order to provide the full picture.
But even then, I cannot log in. I obviously know my user name and password, as I was able to log in from my computer. The error message says "Login error / There was an unexpected error logging in. Please try again. If the problem persists, it may be because you have cookies disabled, and you should check that they are enabled in your browser settings". (It then goes on to explain how to secure my account.)
If you can view logs for troubleshooting, I believe my phone's IP address just now is 178.55.164.227.
Brief update: Cookies was set to "From Visited" in Grazing. Just to check that this isn't the issue, I changed it briefly to "Always", but this did not help.-- era (Talk | History) 11:34, 20 February 2012 (UTC)
Asian font for a punctuation character?
I asked this at the Help desk and was recommended to ask here: At Ellipsis – in Japanese and Chinese, the 3-dot leaders “…” (U+2026) appear on the baseline when the text displays in a Western font. In Asian typesetting it should be vertically centered, and in fact the same codepoint in Japanese (& some Chinese) fonts displays properly centered (like 3 midpoints ···). Is there a template or something with a CSS rule to force display of an East Asian font? I know Unicode obviates most font assignments for world scripts, but in this case it makes a difference visually. I considered substituting the midline ellipsis math operator (⋯ U+22EF) but that would be misleading about the correct character to use. So I thought I should ask for opinions. MJ (t • c) 19:27, 11 February 2012 (UTC)
Thanks, Edokter – {{lang}} sounds like exactly what I wanted. But then I went to edit the section in Ellipsis and found it already used there. I tried changing {lang|ja} to {lang|hani}, but it made no difference. I guess that’s because the ellipsis is in General Punctuation, not an Asian block of Unicode. MJ (t • c) 02:24, 12 February 2012 (UTC)
In my environment (Japnese Windowx XP, IE8), {{lang|ja|…}} displayes properly vertically centered dots (…), while without the template, the dots appear on the baseline (…). It seems the markup is fine and the problem is that the font your browser is using for Japanese is not the one with vertically centered dots for horizontal ellipsis. --Kusunose06:38, 21 February 2012 (UTC)
Random Sampling of AFDs?
Have a look at Template:Recent changes article requests. I'd like to do the same sort of thing, except transcluding the output onto a single page and generating the list of possible titles from the contents of Category:AfD debates. Is there a way that a template could parse that category list and pick random entries, or would we need a bot to generate such a list (in the form of Template:Recent changes article requests/list, for example)? The idea is similar to that of the "Random Article" link, in that you click and get a set of debates you might not ordinarily see. Perhaps articles from outside your usual areas of interest, or in categories you don't normally review. The age of the debates is a factor as well; unless it's the first or last day of the debate, it's difficult to highlight that the article is still up for deletion (apart from the article itself). So this might get more eyes on articles in the middle few days of their debates. Is this feasible? Or am I overthinking it? UltraExactZZSaid~ Did19:46, 13 February 2012 (UTC)
As far as I can tell, it wouldn't be possible to fully automate this type of process using only templates. You can get a list of articles at AfD by using something like {{#categorytree:AfD debates|mode=pages|hideroot}}, but it would be formatted into HTML already, and furthermore there isn't a practical way (that I know of) to just pick a few out of the list randomly and transclude them to a page.
However, this would be a trivial task for a bot to do, and as you indicated on my talk page, the bot I run already works with AfD's and is fully aware of the status of all articles that are currently at AfD. I think this is an idea worthy of a bot task, after some discussion and fleshing out of the details. There are probably several different kinds of such lists that could be created simultaneously. In addition to a list of random open AfD's that is updated a few times per day (perhaps about 25 AfD's on the list), I think it would be useful to maintain a list of AfD's that urgently need more attention from potential voters. The bot could assign an "urgency score" to each AfD and then list the top 25 AfD's by score. I'm envisioning the score being comprised of things like:
The fewer bolded (non-comment) votes an AfD has, the higher its score
The smaller the AfD discussion page is (i.e. in bytes), the higher its score
The shorter the time before the AfD is scheduled to be closed, the higher its score
The more times an AfD has already been relisted, the higher its score
Much simpler than my idea - I like it. You could also check for stagnant debates, using the time since last edit measurement. The trick is coming up with a metric for everything rolled into one score. UltraExactZZSaid~ Did13:18, 14 February 2012 (UTC)
I like it. It's self-evident why the debates are listed (most for lack of participation). Does the bot create a report somewhere that shows actual scores? UltraExactZZSaid~ Did19:17, 16 February 2012 (UTC)
The bot seems to be gravitating toward the most relisted debates, which probably comes from the scoring criteria - so that's working. I think this'll do just fine. Thanks! UltraExactZZSaid~ Did13:26, 21 February 2012 (UTC)
Em 34386| talks about passwords and then changed to username half way through. And WMF would already have details of passwords stored somewhere or you wouldn't be able to log in.EdinburghWanderer21:31, 16 February 2012 (UTC)
No, they don't store passwords - they store encryptions of passwords. During the logging-in process, when you enter a password, it's encrypted and this encrypted version is compared against the encrypted password stored in the database. If these match, the password is assumed to match as well: the chances of two different passwords having the same encrypted form are extremely small. There is no actual means of decrypting the encrypted form. --Redrose64 (talk) 22:31, 16 February 2012 (UTC)
Thats not true - a password hash is one way only --you do *not* need a copy of a password to log in. You need the salted, hashed password. 86.7.36.50 (talk) 22:25, 16 February 2012 (UTC)
That's technically not true either. You need something (not necessarily the original password) that, after the authentication code salts and hashes it, matches the stored password hash. Anomie⚔02:00, 17 February 2012 (UTC)
I'm sorry for the unclear wording. I've clarified it. Wikipedia certainly does not store user passwords in plain text. It uses per-user salted double MD5. It currently has a feature to email you your username and a new temporary password if you forget your current password. You also don't have to use the temporary password, in case it is sent maliciously; it expires in 7 days. The proposed feature is to email this information even if you forget your username. Superm401 - Talk00:22, 21 February 2012 (UTC)
Suggested edit to template and creation of tracking category
{{Rating}} This template accepts values such as {{Rating|9|10}} but not {{Rating|9.1|10}} (as it appears on this revision of Paranoid (album).) Can someone please amend this so that incorrect uses populate a tracking category? I'd be happy to sweep through and fix them. Thanks. Please also respond on my talk if you can to let me know that it's been made. —Justin (koavf)❤T☮C☺M☯ 07:40, 15 February 2012 (UTC)
From the Template Documentation: For example, if a reviewer awards something a rating of "8.6" (expressed in digits) on a scale of ten, you should write it as "8.6" or "8.6/10.0" in your article. Do not use {{Rating|8.6|10}} as it is inaccurate and misleading. (Emphasis is mine. ) It's not a technical issue. Setting it up so that it can show a fraction of a star is technically complex, and probably not worth it. PuppyOnTheRadio (talk) 08:16, 15 February 2012 (UTC)
Indeed. But the documentation does not actually state that fractional numbers should not be used, it states only that the template should not be used for ratings that were not originally expressed in "stars". The template code seems to round any fraction to a half, 1.01, 1.5 and 1.99 would all display identically, like this: , , . Before changing the template, which is edit-protected, perhaps someone familiar with its use could edit its unprotected documentation page with more restrictive guidance, to demonstrate that there is a consensus for prohibiting some (or all?) decimals. Koav might not have realised that the template accepts decimals and displays halves. — Richardguk (talk) 15:13, 15 February 2012 (UTC)
This is a non-issue I think. Reviewers that use stars don't use tenths. The rating in the example should just say "9.1" and shouldn't be using {{rating}} at all (it's actually incorrect now, as the mouseover says 9/10 instead of 9.1/10). As for the tracking category, no opinion, although hopefully it won't be used to round down decimals like on Paranoid. — Bility (talk) 18:44, 15 February 2012 (UTC)
Clarification RedRose is exactly right--I'd like a tracking category like the example given, for exactly the reason that Bility points out: no one assigns "9.1 stars" so any time that appears in an article 1.) the template has been misused and 2.) the template incorrectly rounds these values. It should be amended so that it only accepts "X", "X.5", and "X.0" values and if anyone puts in "X.7" then an error tracking category is generated and someone (me) can come along to fix these errors. —Justin (koavf)❤T☮C☺M☯ 19:14, 15 February 2012 (UTC)
I have limited access to the site at work, so my ability to test is a little limited, but as far as can work out all that it needs is:
<includeonly>{{#ifexpr:{{{1}}} mod 0.5 = 0|<!-- There is no issue. Nothing to see here. Move along -->|[[Category:Pages with incorrect use of Rating template]]}}</includeonly>
Another alternative is to not worry too much about the categorisation, but instead set the template up so that if it picks up the catch I've set up there it dislays it as a numerical figure instead, and otherwise as per the current template set-up. PuppyOnTheRadio (talk) 01:01, 16 February 2012 (UTC)
Really? I could have sworn I've done that in the past with a binary calculator. Stupid parser. An alternative would be {{#ifexpr: {{{1}}}*2 <> floor({{{1}}}*2)..., but either way works. PuppyOnTheRadio (talk) 01:21, 16 February 2012 (UTC)
I don't think that would work for n.5? I ended up using {{#ifexpr:{{{1}}}-floor{{{1}}}<>0and{{{1}}}-floor{{{1}}}<>.5. — Bility (talk) 01:29, 16 February 2012 (UTC)
Well in your example of 0.5 (half a star), 0.5 * 2 <> floor(0.5) * 2 would return true, meaning it would get categorized as being improperly used even though half stars are valid. At any rate, I've tested the Rating/sandbox version in article space and it correctly adds the category or doesn't as it's written. — Bility (talk) 16:14, 16 February 2012 (UTC)
#REDIRECT redirects to an older version if not logged in
I've added a paragraph on RGBA LEDs to LED lamp#Technology overview. It's the third one in this section and it starts with "The color rendering of RGB LEDs, however, is worse than one would expect".
This paragraph doesn't show up if I'm (a) not logged in and (b) am going through a redirect, such as LED bulb. Purging helps only until the next time I clear my browser cache.
I see this on two different machines, one on Win7+Firefox3.6.26 and one on WinXP+Firefox3.6.26, and using two different i'net providers, too.
I've made one more try to get some ops-side explanations on this, then I think it will be time to update WP:Purge and WP:Redirect, even if it's only to say "we're not sure and no-one is saying". Franamax (talk) 21:25, 15 February 2012 (UTC)
Thanks for your help; I've just confirmed with LED bulb that the workaround is to purge the redirect page itself. Another redirect, Led lamp, still delivers an at least two weeks-old version to IP editors.
Parsing the "What links here" page for redirects and purging them looks like a nice job for a bot until the ops get around to fix this problem. --Mkratz (talk) 10:47, 22 February 2012 (UTC)
Would it be technically possible to add a "purge" link to the "Redirected from" message temporarily? Something like "(Redirected from Foo. Out of date copy? Purge the page)" - where "Purge" is a link that will purge the redirect. -- John of Reading (talk) 11:03, 22 February 2012 (UTC)
Looking for Help Gathering Data on Bots
Hello. Some of you might have run into me before...I am doing a research project on bots, bot operators, and technical tools on WP and WM projects. I'm wondering if anyone wants to tackle this problem, which would help me out tremendously. I am looking for stats and data on bots, especially over time. Things like:
(#) of bot accounts registered over time (by month would be fine) (on English WP)
(#) of bot edits over time (on English WP)
(#) of BRFA approved and not approved over time (on English WP only, obviously)
same trends for bot use on other language versions (which would be a bonus)
I've found some info on these things spread around WP, but nothing that is both up-to-date and reasonably accurate/reliable. I'm not sure if getting this info involves dealing with a data dump (I suspect it does), or if there are simpler ways to do it. If you're interested in investigating this with me, I'd really appreciate the help. Please let me know here or on my talk page.
And if you're a bot operator or Wikimedia developer (or someone who deals with the technical infrastructure of WP) and you'd like to be interviewed, please see my call here.
It's very hard to gather meaningful data on bot editing of wikipedia since there are an unknown number of unauthorized ones running from regular user accounts at any time. We just had a messy arb case to get rid of an especially persistent abuser, but there are are constantly new incidents. It's like trying to get statistics on liquor consumption by looking at liquor tax receipts, in a region full of moonshiners. 67.117.145.9 (talk) 21:47, 18 February 2012 (UTC)
Thanks to both of you for the help. Yes, I don't imagine I will be able to get a fully accurate take on bot activity because of those flying under the radar (or the ones who apparently think they are...could you pass along the link to that ArbCom case?). For my project, just getting data on approved bots will be good enough, though. I will check with the people at database reports. Thanks again! UOJComm (talk) 00:35, 19 February 2012 (UTC)
The arb case was WP:Arbitration/Requests/Case/Betacommand 3. That situation wasn't "under the radar", it was blatant, it just took years of drama to make it (probably temporarily) stop. One issue is there's a bogus "legal" definition of a bot (WP:BOTPOL) that enables a lot of wikilawyering over whether something is a bot or not, with lots of programs and editing just skirting the boundaries and avoiding getting treated as bots (even though that is what they are). So the arb case I linked didn't involve long-running, fully-automatic bots (that might be what you're interested in) but a lot of smaller bursts of automatic editing. 67.117.145.9 (talk) 00:53, 19 February 2012 (UTC)
Thank you for the link and the comments. If you would like to be interviewed for my project, I'd love to hear your detailed thoughts. Please send me an email (anonymity can be maintained if desired). UOJComm (talk) 22:49, 20 February 2012 (UTC)
Racist vandalism back again
Could someone with the appropriate technical skills take a look at WP:AN/I, and try to find out why, despite my best efforts, filter 139 has been missing more of the incoming racist template vandalism? Please be very careful viewing the diffs given there: they cover the content with an invisible image that clickjacks every link to racist sites which may well also contain malware, even in preview mode. -- The Anome (talk) 00:42, 19 February 2012 (UTC)
139 is now deactivated, and only 453 is active, and has been tested on defanged vandal template wikitext. Let's see if it catches the next lot. -- The Anome (talk) 01:35, 19 February 2012 (UTC)
I've noted a small issue in the Special:Upload for all logos - the summary field includes {{non-free logo}} at the bottom and the licensing options only include "I don't know..." and non-free logo again; thus most logos now are being uploaded with two copyright templates. Minor glitch, but probably easily fixable. Skier Dude (talk) 02:44, 20 February 2012 (UTC)
I second this. I probably double licensed about 50 county seals in the past week or two before realizing this was why. Not a big deal, but probably fixable, like Skier Dude said. --Andrew(User:90)(talk)20:40, 20 February 2012 (UTC)
Top most 1000 used words
Where is list of 1000 or 2000 top most used words of english wikipedia? Is it possible to do this list if it no.--Kaiyr (talk) 14:09, 20 February 2012 (UTC)
It actually would be nice to generate a concordance of words used in Wikipedia, as this would indicate whether Wikipedia deviates from normal language usage. However, I suspect it would be loaded up with technical terms such as template parameter instructions, and the words "redirect" and "disambiguation". bd2412T16:42, 20 February 2012 (UTC)
A possible topic for research, I'm sure, but not really appropriate for encyclopedia writers. Just sayin'. --Izno (talk) 18:46, 20 February 2012 (UTC)
Sort challenge
In User:Sphilbrick/List of current NCAA Division I women's basketball coaches, the table is set up to be sortable. I have a subheading in the first column so that one can edit a row, without having to edit the whole table. That adds a numbering to the entries, which is fine. However, if I sort on the team column, it sort such that 1 is first, then 10, rather than 2.
I tried adding <span style="display:none;">Air Force Falcons</span> inside the headings, but that didn't seem to help; my guess is that it converts the heading code to numbers, then sorts. I tried adding the span style outside the headings, but then it doesn't render the headings.--SPhilbrick(Talk)16:05, 20 February 2012 (UTC)
What numbering? I don't see any numbers and it sorts fine for me. I went through all your revisions and never saw any numbers. Which skin are you using? — Bility (talk) 16:13, 20 February 2012 (UTC)
I turned on section numbering and tested some invisible spans in the table and it seems to work for me. It look all right on your end? — Bility (talk) 16:31, 20 February 2012 (UTC)
@PrimeHunter Thanks, I've had autonumbering on for so long, I forgot it wasn't a default.
@Bilty I was trying to imbed the invisible spans in the team entry, as oppose to a new line. I see you added through Bradley, is there an automated way to do this, or is it one at a time? In any event, thanks for identifying the problem.--SPhilbrick(Talk)16:52, 20 February 2012 (UTC)
Is the list going to grow? You might want to make a template for these entries in your userspace and use that to cut down on clutter and make it easier to add teams. — Bility (talk) 17:25, 20 February 2012 (UTC)
No, the list is essentially complete (one or two new teams a year show up, but that's pretty minor), although I need to add the sort key to every row, so wondering if that can be done in an organized way rather than one at a time. When you did the first 20 or so, did you do each row manually, or did you have a better technique?--SPhilbrick(Talk)22:48, 20 February 2012 (UTC)
Did them manually by copy/pasting "<span style="display:none;"></span>" then highlighted the team name and Ctrl-dragged it into the span. If you want, I can do the list fairly quickly with some regex. Do you want to keep the span or make a template? — Bility (talk) 23:00, 20 February 2012 (UTC)
I also need to add the rest of the headings, plus I need to add a sort key for the name, so they will sort by last name , not first. Let me try dumping it to Excel and trying to fix it myself, first. --SPhilbrick(Talk)23:31, 20 February 2012 (UTC)
If you see no difference, it's possible that the font (in this case Lucida console) isn't installed on your machine. When this is the case, most browsers will fail gracefully, by displaying a fallback font. The example below <span style="font-family:'Courier new',monospace">Courier new, with monospace as fallback</span> shows how this may be amended: if Courier new isn't installed, it'll try monospace; if that isn't installed either, it'll default to the page style. --Redrose64 (talk) 21:31, 20 February 2012 (UTC)
(edit conflict) Sure, just look further up this page at the technicolor riot. There are two main ways: using <font face="xxx">...</font> or <span style="font-family:xxx;">...</span>, where xxx is the font name. Be careful that use of non-default fonts does not violate WP:ACCESS though. I've spotted the following (colours removed for clarity): arialCalibriCopperplate Gothic LightCourier new, with monospace as fallbackpapyrusTahomaTrebuchet MSTrebuchet MS (done another way) --Redrose64 (talk) 21:26, 20 February 2012 (UTC)
All MiszaBots have been down for three days now. User:Misza13 has not edited since last May and cannot be reached via email, so if anyone has any other ways to reach him/her, can you please do so? Talk pages and noticeboards are starting to pile up and require manual archiving at this time. --Andrew(User:90)(talk)22:20, 20 February 2012 (UTC)
Some people believe that nightshade.toolserver.org is down, and that's the reason why the archive bot is not running. Sending a ping to 91.198.174.201 gets 'destination host unreachable'. If the nightshade server is going to be down for a while, maybe someone could move the crontab from there to one of the working toolserver machines. EdJohnston (talk) 03:05, 21 February 2012 (UTC)
I don't know why you say "some people" believe. Nightshade is down and bot operators have been advised afaik to move their cronjobs to willow, which is still operational. As I understand it, the issue has persisted since the extraordinary maintenance on Toolserver on the 16th and 17th, but I might be mistaken on this. SnowolfHow can I help?07:10, 21 February 2012 (UTC)
Misza has responded at User talk:Misza13#Re: Bots down & stuff. He indicates he will run his archive bots manually beginning at 16:00 on 21 Feb. until somebody fixes nightshade. (He thinks moving the cron jobs to willow could create further problems). The nightshade machine has been down since Friday 17 Feb. In the toolserver issue report, DaB has stated: "Looks like I disconnected the disc-controller or the disc by accident. Needs someone in the colo to fix this." EdJohnston (talk) 14:23, 21 February 2012 (UTC)
Unable to enter text at one site
I wanted to leave a note at talk:Inflation rate world.PNG that the map is in error (It shows the US with 0% inflation, whereas the cited source gives the US's inflation rate as 3%; I didn't check any other countries), but when I twice tried to leave a comment, my heading got erased and I could not (physically) enter any text at all into the text box. Kdammers (talk) 06:02, 21 February 2012 (UTC)
A query made up of two or more words produced no results until 08:45 of the 21st of February 2012 (UTC). It now runs fine in the main namespace, but not in other namespaces, where it produces no results. Happy editing! –pjoef (talk • contribs) 11:23, 21 February 2012 (UTC)
It looks like "/wiki/Special:Search?" works but "/w/index.php?title=Special%3ASearch&" doesn't. Simply replacing this part of the URL fixes the search. I think it that the "?" and the "&" in the query strings and not placed in the correct places when used the second form.
I hope you plan on getting a wider consensus. I don't think the few people watching WP:PORTAL should be changing something as far-reaching as infoboxes. — Bility (talk) 15:58, 21 February 2012 (UTC)
Well, it was an RfC, but yes, I understand. Hence I'm asking for someone to sandbox this, just so we have some idea what it will look like. FWIW, I couldn't care less where the Portal links go, but since some people thought it was a good idea I figure there's no harm in testing this. NoleloverTalk·Contribs16:04, 21 February 2012 (UTC)
Your sandbox is last? Mine is third (Redrose64 My talk My sandbox My preferences), so I guess the position isn't set in stone. I have no special code for setting it - it's just the default position when I have the gadget 'Add a "my sandbox" link to the personal toolbar area.' turned on. --Redrose64 (talk) 18:22, 21 February 2012 (UTC)
Alex' solution works. Now 'Log out' is at the far right side and 'My Sandbox' is directly to the left of it. Thanks very much to all for the help. Toshio Yamaguchi (talk) 18:34, 21 February 2012 (UTC)
Clicking on External Link sets off antivirus alert
I was editing Saint Thyrsus and one of the External Links was to www.saintpatrickdc.org etc. When I clicked on the link it set off my antivirus program (ESET NOD32). This has never happened to me before in years of editing. I deleted 2 instances of this link (I was making the references look better), so it's not on the current article page—you'll have to look at older versions of the article. This is the message from my antivirus program: ESET NOD32 Antivirus - Alert
Access denied !
I found a way of fixing the problem. Instead of a URL that ends, for example, with .htm#thyr, if I change it to .shtml, it appears to take me where I want to go. (I can drop the #thyr because the original didn't take me to that section of the page anyway.) An "everything" Search for www.saintpatrickdc.org yields 463 results, almost all of them Catholic saints. It looks like I've been given a project for Lent. Thanks for your help! --Kenatipospeak!00:45, 22 February 2012 (UTC)
I've opened several of the old .htm URLs now and the worst that happens is a "404 page not found" message. My antivirus program may have erred on the side of safety initially, but that's OK with me. --Kenatipospeak!21:02, 22 February 2012 (UTC)
How to add a parameter to a template?
I designed User:Piotrus/TDYK for notifying people about DYKs. I'd like for it to be more specific; I'd like to be able to use a {{subst:User:Piotrus/TDYK|parameter - article's name}}, so that the template would actually mention the name of the article that I specify in the parameter. How can this be done? Perhaps somebody who understands the templates better than me could edit my template and add that functionality to it? In similar vein, I'd like my User:Piotrus/w welcome template to be responsive to a parameter that would specify the name of a WikiProject that the person I am inviting may be interested in joining. --Piotr Konieczny aka Prokonsul Piotrus| talk to me23:22, 21 February 2012 (UTC)
Where is all the JavaScript for the collapsible boxes on Commons? Specifically, with JavaScript I would like put a hook that executes when the user clicks the box for dropdown. Magog the Ogre (talk) 23:46, 21 February 2012 (UTC)
I got the "Sorry! This site is experiencing technical difficulties" notice twice a minute ago, but my edits still went through. Was this just a hiccup, or are there things going on? Ten Pound Hammer • (What did I screw up now?)01:12, 22 February 2012 (UTC)
I'm looking for a solution for someone which will switch page layout from a 'fluid grid style' design to one which adds a max-width statement, to improve legibility for those with difficulty reading long lines. Is there such a tweak, either in .js or .css that I can recommend? Ocaasit | c05:04, 22 February 2012 (UTC)
Add
#bodyContent{width:80%}
to your skin.css and tweak the percentage as necessary. Note that this shortens the width of all pages, including watchlist and contribs pages. You can disable this effect on special pages by prepending .ns-special to the css above (.ns-special #bodyContent...) adding another line of code: .ns-special #bodyContent { width:100% }. Goodvac (talk) 05:25, 22 February 2012 (UTC)
On WP:TFD, Wikipedia:Templates for discussion/Log/2012 February 16 and Wikipedia:Templates for discussion/Log/2012 February 15 are just showing up as plain links instead of transclusions of those day's logs. Is this just due to template overflow? Ten Pound Hammer • (What did I screw up now?)21:18, 22 February 2012 (UTC)
This probably isn't even an issue for VPT, but it seemed like a place I could get a quick answer; I've always used Firefox, where my talk page looks fine with the discussions left-aligned, but in *cough* Internet Explorer the discussions are centred, which looks terrible. Could someone tell me what the difference is that's causing the problem? Thanks. Black Kite (talk) 01:45, 23 February 2012 (UTC)
In IE, table cells do not inherit the text-align property from the table. And by default, <th> tags are center-aligned. Apply the alignment directly to the cells instead of to the table, and it should work. Anomie⚔01:57, 23 February 2012 (UTC)
Also/or: use "|" instead of "!" at the beginning of the last line of table code, to make it a normal left-aligned <td>...</td> element instead of a bold centered <th>...</th>. In any case, having so many unclosed tables and unclosed table cells is taking chances with the wikitext parser! — Richardguk (talk) 02:06, 23 February 2012 (UTC)
This may be impossible, but is there a way to transclude a template into a page without that page ending up on a category in that template? (Assuming the template is generally designed to have the page added to a category.) PuppyOnTheRadiotalk12:29, 23 February 2012 (UTC)
Firefox 10.0.2: The wiki structure is disrupted. I can't see tabs and the left panel. Images do not have default look. Is anybody else experiencing the same??? RedtigerxyzTalk16:47, 23 February 2012 (UTC)
I can see the left panel in Commons, but the Commons logo on the home page is not visible. Except that everything else looks good in Commons. RedtigerxyzTalk16:50, 23 February 2012 (UTC)
Does this page start with:
The Wikimedia Foundation's 2012 steward election has started. Please vote.
More than 50,000 articles: Bahasa Melayu · Български · Eesti · Ελληνικά · Simple English · Euskara · Galego · עברית · Hrvatski · Lietuvių · Norsk (nynorsk) · Slovenčina · Slovenščina · Srpskohrvatski / Српскохрватски · ไทย
...
I can't see "The Wikimedia Foundation's 2012 steward election has started" banner anywhere on the page in Mozilla, but can see it in IE. RedtigerxyzTalk17:06, 23 February 2012 (UTC)
Sounds like either the stylesheets aren't loading or you have CSS disabled. Do you have Firefox Web Developer installed? Check Firefox View (you may have to press ALT to show the menus) → PAge style. ---— Gadget850 (Ed)talk17:46, 23 February 2012 (UTC)
When I post that link on my Facebook page, it should display this preview:
I was born a poor farmer in rural India in 1936. Today I rely on and edit Wikipedia.
I want Wikipedia to be here for all future generations. This is our annual fundraising drive to pay for the servers, small staff and other infrastructure that keeps Wikipedia on the web for free...
But instead, the preview displays the following:
$( document ).ready( function () { // Disable submitting form with return key $( 'form' ).bind( 'keypress', function(e) { var code = ( e.keyCode ? e.keyCode : e.which ); if ( code == 13 ) return false; } ); } ); /* This CSS is responsible for the overall layout of the LP*/ #LP-table {…
Thanks for filing the bug. This is definitely on our radar. We are reworking the entire landing page system for the next fundraiser and hope to fix this and a multitude of other issues in the process. Following bug 34660 is probably the best way to get updates on this. Pgehres (WMF) (talk) 10:40, 25 February 2012 (UTC)
Custom drop down options on special pages
How can I customize personally options found at MediaWiki:Ipbreason-dropdown, MediaWiki:Protect-dropdown, MediaWiki:Deletereason-dropdown and other drop down lists found on various administrative special pages? For example, how could I make "Repeatedly removing speedy deletion templates from own created pages" appear in the Ipbreason-dropdown list for me without actually adding it to Ipbreason-dropdown? I had it suggested that such customization could be done via javascript or something similar, but I'm not sure if this is even possible, let alone how to do it. Any suggestions? Ks0stm(T•C•G•E)18:29, 23 February 2012 (UTC)
I tried throwing together some javascript to add additional reasons to the block form (easy enough), but apparently there is some form sanitation code in the MediaWiki engine that blocks any submitted values that aren't on the block reason. Reaper Eternal (talk) 20:03, 23 February 2012 (UTC)
You could make it so when you click a custom reason, it pre-fills the 'mw-input-wpReason-other' textbox then changes the reason to Other. That should be useful and will not be detectable by the server. Superm401 - Talk00:40, 24 February 2012 (UTC)
I'm a dark background kind-a-guy; I set my browser to ANYTHING but white ... so equations, formulas, practically anything notated at wiki appearing as clear-backgrounded image is absolutely unintelligable
My suggestion is that since those browsers who leave their settings with light background would be unaffected by a change from CLEAR to WHITE backgrounds in these script images, why not make EVERY equation/etc white backgrounded?
That way, despite my dark background, I (capital) would be able to read the darn things. — Preceding unsigned comment added by 67.102.65.21 (talk • contribs) 19:23, 23 February 2012
Are you using either the background/foreground color selection in the web browser options or the operating system's High Contrast mode to force the black background? Either way, you would only see a white outline around the black text. If I am not mistaken, Wikipedia will, sooner or later, switch to rendering these formulas in JavaScript using MathJax (so there would be no images involved). I know of no good workaround, as most CSS changes do not affect High Contrast mode.
If you happen to use Internet Explorer as your web browser, you could create an account here, log in, and then go to Special:MyPage/common.css and insert the following code to make the formulas white-on-black. Then press Ctrl-F5 to load the new settings.
Basically you can access all of soxred93's tools by replacing the soxred93 part of the URL with tparis. The user that previously operated the tools has retired and TParis now operates them. --Andrew(User:90)(talk)06:06, 24 February 2012 (UTC)
The intention is to link a version of the respective page at a currently interesting point in time (e.g. last page version in february 2007, maybe inclusive more accurate parameters) or, similar to FullURL and #expr, to link a version of the respective page [xy] time ago. --Hæggis (talk) 17:42, 24 February 2012 (UTC)
It's not moved over to Wikipedia. You can tell that it's still on commons because of the presence of this box below the resolution information. Commons images are accessible on Wikipedia under the same names - the source is transparent. --Redrose64 (talk) 21:38, 24 February 2012 (UTC)
I don't get what you mean. When I originally linked it, it was on the Commons. I found it by looking through a list of images on the Commons that hadn't been categorized yet. But now when I click the link, its on wikipedia. I can't seem to get back to the Commons version. MathewTownsend (talk) 21:50, 24 February 2012 (UTC)
Yeah, but if I just click the link, it goes to a wikipedia page - even though I linked it to the Commons page. Am I supposed to move it over? Other times I've added an image from the Commons, there hasn't been a problem. MathewTownsend (talk) 22:12, 24 February 2012 (UTC)
The scaled-down versions of the tif won't load for me either. But the full size tif is a ridiculous 8 Mbytes. Have you tried converting to a PNG? — RHaworth (talk·contribs) 22:23, 24 February 2012 (UTC)
I thought that I explained... the link goes to something that only looks like a Wikipedia page. It is a "page" that doesn't actually exist. The image is on Commons. Its description page is on Commons, see Help:File page. Those are the proper places. Nothing needs to be moved anywhere. The whole point about having images on commons is that they are available to be shared by all Wikimedia projects, without having to make local copies on all the different Wikipedias, see Wikipedia:Wikimedia Commons. Images on commons are placed in articles, or linked, using exactly the same syntax as for images held on Wikipedia. There is no special technique, nothing that you can do to the image linking code that means "I want this image to come from Wikipedia", or "I want this image to come from Commons". When you make an image link, it looks for an image of that name on Wikipedia and presents that if available. If there isn't one, it looks on commons and if there is one there, returns the image with the same appearance that it would have had it been held on Wikipedia with the exception that the box MediaWiki:Sharedupload-desc-here is contained within the page. --Redrose64 (talk) 23:00, 24 February 2012 (UTC)
The page Help:Contents uses an inputbox with "namespaces=Help**,Wikipedia**,Template**,Category". According to mw:Extension:InputBox, this means the first three checkboxes should be ticked by default. An IP has noted here that this isn't working - by default, no checkboxes are ticked and only the main namespace is searched. -- John of Reading (talk) 22:09, 24 February 2012 (UTC)
I don't know why but a comparison [5] of produced code for the above inputbox with namespaces=Main**,Help shows that here at en.wikipedia.org a parameter checked="checked" is not made when a namespace has ** after it. PrimeHunter (talk) 22:50, 24 February 2012 (UTC)
I receive a steady stream of e-mails (via the 'E-mail this user' link) on matters which should be dealt with on talk pages. I want to give people who do this a prompt, automatic message about it.
I created a special GMail account and set up an Out of Office AutoReply (GMail is schizophrenic and also calls it a Vacation responder) to give a message saying "if it is a Wikipedia matter, use my talk page".
Seemed to be a good idea - one problem: GMail is sending the AutoReply to the bounce address, wiki<at>wikimedia.org and not to the 'From' address of the sender. Any suggestions?
(Ironically, the first e-mail I received on this new accout was from someone who I had blocked with 'cannot edit own user_talk page' - so the auto message would have been pointless.) — RHaworth (talk·contribs) 22:19, 24 February 2012 (UTC)
Problem displaying panoramic images (Android browser)
Hi, in the Android browser (tablet), Wikipedia pages with very wide images, usually panoramas, are scaled so that the whole image fits the width of the screen, resulting in text so small that it is almost illegible. I wondered if there was anything that could be done at the Wikipedia end to stop this happening.* I know that browser compatibility issues are a pain in the neck... 81.159.106.110 (talk) 04:19, 25 February 2012 (UTC) *Or at the browser end, should anyone happen to know!
User last edit
Hi. I'm trying to generate a list of users who are in a particular user group (ie Admins) that also reflects the date/time of their last edit, so as to work out what admin is on the site at any given time. I'm assuming that there is a way to do this using dpl, but that is unfortunately not within the list of my coding skills. Anyone have any options? PuppyOnTheRadiotalk08:32, 25 February 2012 (UTC)
Please see the "MediaWiki:Asksqltext — delete or to be updated?" section of WP:AN. Billinghurst and I think that the page should be deleted, but we're not sure what it does/did or what effects (if any) its deletion would have. Nyttend (talk) 03:24, 26 February 2012 (UTC)
Redirects to Wiktionary
Normally, soft redirects to Wiktionary are excluded from being counted as articles by technical services such as the Untagged Uncategorized Articles tool. However, what seems to happen is that if somebody replaces such a redirect with a dicdef article, and then somebody else reverts that back to the redirect again, then the page suddenly does start getting picked up as an uncategorized article, and won't subsequently drop from the list through most normal means. To date, the only workaround that has proven effective at all is to add the redirect to Category:Temporary maintenance holdings, a category I created specifically as a crapcatcher for random oddball situations like this — but that isn't a viable permanent solution to the problem.
Does anybody know what else can be done to fix the technical issue here, and cause the system to properly revert back to recognizing the page as a Wiktionary redirect again so that it doesn't get picked up by article-scanning tools? Thanks. Bearcat (talk) 19:08, 26 February 2012 (UTC)
Without knowing how this Untagged Uncategorized Articles tool works, it's unlikely that anyone can guess what the problem is in order to find the solution. You might be best served by asking the author of the tool directly. Anomie⚔03:07, 27 February 2012 (UTC)
According to the time frame you have there, the drop occurs about 3 or 4 hours after the Signpost's special report about page patrolling. We were looking at this today, actually, and have concluded that it was a series of inactive patrollers "waking up" and clearing the queue, thought that may be an incorrect assumption.--Jorm (WMF) (talk) 01:19, 23 February 2012 (UTC)
I suppose the patrol log would reveal the answer, if one was motivated enough to look through it. 11 days of articles is a lot of patrolling for a few patrollers to do though. Kinda makes you wonder if someone just opened up 1000 articles and hit the "patrolled" link without even looking at them. —SW—express16:30, 23 February 2012 (UTC)
I did a little digging and it's not adding up. According to an SQL query on toolserver, only 81 articles were marked as patrolled on February 21st between 19:00 and 20:00 UTC (that's 81 articles in article space only, which were manually marked as patrolled). Yet, the graph shows that the newpages queue lost 6 days worth of articles in that hour. How is that possible? —SW—yak23:38, 23 February 2012 (UTC)
Nope, my tasks run on willow. And besides, the newpages queue hasn't been in the single-digits since... well, the last time that mediawiki software was updated and the entire queue got wiped out. The only logical explanation I can think of is that there were a bunch of editors who went on a patrolling spree in the middle of the queue, over a few days, and then when the back of the queue caught up to the hole they made in the middle, there was a big discontinuity in the graph (since the tool which creates the graph only looks at how old the last few articles in the queue are). But still, to erase 11+ days worth of articles from the queue is a pretty monumental task, unless you're barely looking at each article, which is not the point of NPP. —SW—spill the beans15:20, 24 February 2012 (UTC)
I'm sure this has been asked before, but how do I keep red links from going directly to edit mode? This is more than useless to me and incredibly annoying. —danhash (talk) 20:16, 24 February 2012 (UTC)
One option is to get yourself blocked. Or you could try this userscript to "fix" the red links:
//red links not go to edit mode$(function(){$('a.new').each(function(i,aa){aa.href=aa.href.replace(/[?&]action=edit/,'')})})
I am usually logged-in while using Wikipedia (and for good reason—I edit), so logging out does not fix my problem. —danhash (talk) 21:25, 24 February 2012 (UTC)
Thanks Alex, I'll try it out. I really don't need more JavaScript than is necessary executing on every page load though; I already have enough user scripts that could/should be MediaWiki functions as it is. If I were a casual user it wouldn't be that big of a deal, but especially when I'm working on long pages from moderately-powered computers it gets pretty ridiculous. If MediaWiki can change it for blocked users there should be an option in preferences for everyone. Is a user script the only current way? If so, I propose an option be added by the devs; should be pretty simple. —danhash (talk) 21:04, 24 February 2012 (UTC)
&redlink=1 tells MediaWiki not to automatically open edit mode if you do not have permissions. I'm not sure if there are any other uses. — AlexSm22:29, 24 February 2012 (UTC)
I can't think of any good reason for the default action for clicking a red link to be an edit window. Most of the time (at least for me) clicking a red link is a way to get information about the page, such as deletion logs or "what links here", not to edit the page. If I want to create the page I'll simply click "create". Can someone add the ability to MediaWiki for red links not to go to edit mode? The JavaScript above is useful when it works, but there are places where red links aren't modified not to go to edit mode; plus, nobody needs more JavaScript than is necessary loading on every page. —danhash (talk) 21:53, 27 February 2012 (UTC)
This image File:Zacuelu1.jpg is displaying incorrectly as a thumbnail on the Zaculeu page, with severe colour distortion when using Firefox. The commons file looks fine, and it displays fine on internet explorer and opera. Clicking on the thumbnail produces a normal-looking photo. I remember there being a bunch of photos like this on the page some time ago but can't remember how they got sorted out. Any help greatly appreciated, Simon Burchell (talk) 01:32, 26 February 2012 (UTC)
For me, it shows just fine in IE7, Chrome, Safari and Opera, but in Firefox 3.6.27 it shows as a monochrome image: black and magenta at all resolutions. It's not a full-colour image from which the green has been removed - there are no reds or blues, just shades taken from a magenta scale like this:
The image has an embedded color profile ("SFprofT (OpticFilm 7600i).icc") that Firefox may be confused by. Converting the image to default RGB may fix the problem. — Edokter (talk) — 15:14, 26 February 2012 (UTC)
Open the original image in Gimp, then either assign the default RGB profile (to show the original colors), or convert the embedded profile to RGB (to show the intended effect), save the file and reupload to Commons. — Edokter (talk) — 15:52, 26 February 2012 (UTC)
Thanks everyone who replied - I've converted the colour profile in GIMP as suggested and uploaded the corrected version, everything looks fine. All the best, Simon Burchell (talk) 18:26, 26 February 2012 (UTC)
This is one of those features that's been lacking from HTML right from the start (well, since HTML 3.2 anyway which is when tables were added). The thing is, there's an enclosure for the cells making up a row - this is the <tr>...</tr> element - to which attributes may be applied. There is no enclosure for a single column - the next enclosure outwards from <tr>...</tr> is <table>...</table>; again, attributes may be applied to that, but they affect the whole table. But, certain attributes do have an effect on a whole column - the width is one of them, so if you put style="width:10%;" into the attributes of any cell, that whole column will be forced to that width. --Redrose64 (talk) 20:03, 27 February 2012 (UTC)
Script that fetches references mentioned in a section I'm editing that aren't currently available?
Let's say I'm editing a section, and there are references cited in it (such as <ref name="example" />), but the ref itself is declared in another section. Anyone know of a script that fetches these automatically while editing a section so I can see the ref's info? Surely this is something that would be fairly popular if it existed or was widely known? I looked through WP:JS and don't see anything there that does this.
Lately I've been using the References Segregator to move all references out of inline and into the References section so that text is less cluttered, and with a script like the one I'm looking for, I wouldn't need to keep a window open of all existing references while editing. Gary King(talk · scripts)21:58, 27 February 2012 (UTC)
Unfortunately I'm not a big fan of wikEd as it's too bloated for my taste. I like some of its components, though, like the one you mentinoed, as well as wikEdDiff in particular.
EDIT: I just checked wikEd, and it doesn't look like it retrieves refs not in the current section, both when editing and previewing a section.
Oh, I just recalled the Ajax Preview script, which essentially does what I'm asking. You click a button while in Preview and it fetches the refs that are currently cited. There are a few things that I'd like to change from the script, such as making it run immediately rather than require a button click, and change the loading symbol to show "Loadinfg..." instead, as well as stopping the script from modifying existing buttons. My own version of the script is at User:Gary King/ajax preview.js, so take your pick. Gary King(talk · scripts)22:09, 27 February 2012 (UTC)
I used to consult an application site that counted how many hits per month a Wikipedia entry received. The site was http://stats.grok.se/, but is not longer operating. I wondered if a reader of this entry could direct me to another site that could do the same thing. Thank you. Iss246 (talk) 22:49, 27 February 2012 (UTC)
So I created that page now, but the link still goes to Wikipedia:Upload. Is it just a caching issue, or did I do something wrong? Nyttend (talk) 02:30, 29 February 2012 (UTC)
commons:MediaWiki:Upload-url controls the URL for the upload item in the "participate" section in the sidebar; the upload link in the toolbox on Commons is hidden using CSS. The only way to change the upload link in the toolbox (besides poking at it with Javascript) is to have a sysadmin set $wgUploadNavigationUrl in the wiki configuration (as mentioned above), to install an extension that will change it in the BaseTemplateToolbox hook, or otherwise modify the MediaWiki software. Anomie⚔02:49, 29 February 2012 (UTC)
Short term: add ‎ right after the template. This invisible formatting character says "from here on, L-to-R letters". Without, brakets &txc are handled as belonging to the hebrew, R-to-L text. -DePiep (talk) 12:00, 28 February 2012 (UTC)
Comprehensive blocking of banners, ads, and other intrusive announcements
What does it take to get rid of ALL banners, banner-like messages, ads, text announcements, and other intrusive annoyances? A few days ago when I logged into Wikipedia I saw yet another extraordinarily annoying message at the top of my watchlist page ("Are you interested in encouraging more university professors to use Wikipedia as a teaching tool? Do you enjoy leading, coordinating, and organizing teams? If so, we want you! We are now recruiting for Wikipedia Regional Ambassadors to the multiple regions of the U.S. Click here for more details. [hide]"). Big bold grey text that even after hidden leaves a huge chunk of whitespace at the top of the watchlist until the page is refreshed and that comes back every time you log in. I am tired of this shit. I would rather have external ads on Wikipedia, because those are generally extremely easy to block. The SOPA blackout was about an issue that legitimately affected all Wikipedia users; this banner is not. I don't care who is recruiting for what, what is being voted on, what ArbCom is doing, or anything else that I have ever seen a banner about enough for it to intrude on my screen real estate. Many of these banners are also dynamically loaded with JavaScript so that the page fully loads without them and then they pop up, skewing the whole page that has already loaded and pushing all the content down an inch or more on the screen. Or else they are loaded statically and only dynamically hidden after the page loads. I have WP:Goings-on watchlisted and I would be happy to watchlist any other page that has current important announcements. I would love to have a "Current announcements" link on the left menu that I could click on when I want to see what's new with Wikipedia and the WMF. But I do NOT want any more crap taking up space on my screen, especially when it behaves like it was written by an incompetent script kiddie who once read "HTML for Dummies" in 4th grade. My common.css hides sitenotices and Wikipedia ads. I already had to update the blocking code for Wikipedia ads once when it stopped working. I tried adding CSS to block the new watchlist message, but it ended up blanking most of the page and I don't have the time or patience to troubleshoot it. From all the times that I have searched and searched to find an answer to the question of how I can get rid of these intrusions and annoyances I have not been able to come up with a comprehensive answer. I don't want a userscript to fix this; JavaScript is dynamic, and changing the page after it has loaded is an unacceptable solution. The only acceptable solution is either: 1. CSS code to add to my common.css that blocks ALL such messages permanently i.e. something I don't have to update every time somebody adds a banner in a new place; or 2. preferably just give me an option in preferences to turn this shit off across all WMF sites. Then a page in the Wikipedia: namespace should be created to inform other users how to block these annoyances; I am perfectly willing to create the page myself. Frankly I am surprised that this issue has not been resolved before now. —danhash (talk) 15:37, 28 February 2012 (UTC)
#siteNotice , #watchlist-message, .geonotice{ display: none !important; } would hide both forms of watchlist notice (the ones above, which are watchlist-geonotices, and the ones below, which are plain ol' watchlist notices. I shall leave you to your conclusions as to their merits. - Jarry1250[Deliberationneeded]15:52, 28 February 2012 (UTC)
I've added #siteNotice to my previous message, which covers the ones that display on every page (including fundraising I think). Can't think of any other notices off the top of my head. - Jarry1250[Deliberationneeded]16:51, 28 February 2012 (UTC)
I keep seeing them in more and more places over time. I see them on multiple WMF wikis as well, so maybe it's higher up than administration here. Anybody know who is in charge of annoying users? —danhash (talk) 15:45, 1 March 2012 (UTC)
An article that uses tables with lots of text in individual cells?
Hey,
Does anyone happen to know an article that deals elegantly with table cell fields that have greater than usual amounts of text?
I'm trying to learn advanced table formatting and it would be very helpful if someone could link to an article that deals with this problem: when a table is the best data structure to use, but a few cells need to have large amounts of text in them.
How do you mean? Every cell has text? One particular column ("Notes" as a header)? I know that the first is, if not forbidden, then highly discouraged due to web semantics, while the second is commonplace. Or is there something else you're trying to figure out? --Izno (talk) 16:41, 28 February 2012 (UTC)
Thanks for replying. I mean, a table where a particular cell has a great deal of text. For example, a cell in a table that had three hundred words of text. I'm wondering if there's a way to handle this kind of thing elegantly and was looking for examples. Thanks!
{{Episode list}} places the episode summary in a table-wide cell, and all other info in a multi-cell row above it. See example at Doctor Who (series 5)#List of episodes. If everything was in one row then it would give poor formatting with lots of blank space when the summary often has far more content than the other cells. The summary parameter is documented at {{Episode list}} with "ShortSummary (optional) A short 100-300 word summary of the episode". I find the name ShortSummary a bit ironic in that context but some fans tend to write very long summaries. I guess the name is intended to restrain them. PrimeHunter (talk) 23:36, 28 February 2012 (UTC)
I don't mean for this to sound like an overly snide comment but if you don't want IP's to edit then you should just mandate that you must have a username to edit. Requiring 3 or more capthas for an IP to make one change is senseless and a pretty strong message that you would rather not have IP's editing content. Every time I try and edit as an IP it takes no less than 3 captchas to do one change. 71.163.243.232 (talk) 00:43, 29 February 2012 (UTC)
If I recall correctly, you only have to answer a CAPTCHA if the edit you are making involves an external link. See WP:IP. You may have had to answer 3 if you didn't enter it correctly the first two times. Killiondude (talk) 01:52, 29 February 2012 (UTC)
Then something must not be working. It will let you edit a talk page without a captcha but if you make any changes to any pages it has required at least 3 and as many as 5 before I stopped trying. Try and add an infobox or a portal as an IP. 71.163.243.232 (talk) 02:12, 29 February 2012 (UTC)
I just used IE's InPrivate Browsing to add an info box to a random page (that could use one) while logged out and it worked with no CAPTCHA (it did not include an external link): [6].
I have also adjusted an external link, including adding one, which did require exactly one CAPTCHA, even though I made a whitespace edit to the article displayed when the CAPTCHA was requested: [7] (I then made a further edit to that page adding {{it}} twice and was not ask for a CAPTCHA even though it was the same line the external links were on.)
Finally I still intend to test adding an infobox with an external link, to exactly test User:71.163.243.232's complaint, but unless I give in and use a sandbox, I might take a while locating a good page to edit.Mark Hurd (talk) 07:07, 29 February 2012 (UTC)
I added an infobox person with an external link, admittedly one already on the page, and I was not asked for a CAPTCHA again (same session where I was earlier asked for a CAPTCHA once): [8]Mark Hurd (talk) 07:39, 29 February 2012 (UTC)
After looking at 71.163.243.23's contribs, I tried some tests in the sandbox. No captcha when making an ordinary edit but it did ask me for a captcha when using the "undo" function. On second thought maybe this "feature" should be enabled for logged in users too. (but not for rollback) Might slow down some edit wars :) The last part was semi sarcastic. --Ron Ritzman (talk) 14:22, 29 February 2012 (UTC)
Intermittent search problems
Off and on, when I try to search, I get no results. Redlinks just say "there is no article" without the usual list of articles in which the search string or something similar appears, and when I search for articles I know exists, it just says "there is a page named..." without listing anything at all below it. This happens daily, off and on, sometimes alternating between this and "normal" (actually getting results) from search to search with just a few seconds in between. Server load? - The BushrangerOne ping only06:55, 29 February 2012 (UTC)
I've seen this too. In my case it is normally the first search for something, and I think I've seen more than zero results, but on clicking search again, without changing what namespaces are included, you get the many results expected. Sometimes it seems the first search results do not include fuzzy results. Mark Hurd (talk) 07:11, 29 February 2012 (UTC)
The Search index is not updating properly - spelling mistakes I corrected on 25 Feb are still appearing in searches, whilst hitting refresh generates different selections of results. The usual reason is "a stale version of one of search index slices" whatever that means. Could someone please contact whoever deals with this problem - it is making us WikiGnome's work very difficult. Arjayay (talk) 19:23, 29 February 2012 (UTC)
I noticed this on Portuguese Wikipedia/ Wikibooks some hours ago. At some point I was getting results in one of the following pages and not on the other:
I don't seem to be able to preview or view the changes of my edits in the mainspace. I can save pages, but it page hangs and does not get loaded afterward. I do not have the problem in projectspace. It does not seem to matter if I'm logged in or whether I use Chrome or IE8. Any ideas as to what is wrong? —Ost (talk) 15:30, 29 February 2012 (UTC)
This problem went away for me but lingered when I tried to preview or view changes through some toolserver apps. Now it seems to be completely resolved, though I don't know its cause or solution. —Ost (talk) 19:09, 29 February 2012 (UTC)
The database has spent a good bit of time in read only mode lately, so that may have been the problem. I have had it happen to me a couple times in the past week or so. --Andrew(User:90)(talk)19:16, 29 February 2012 (UTC)
Error with WP:AFT v4 on Halo 3 OST
I'm having a visual error pop up and I'm not sure why. On Halo_3_Original_Soundtrack, I'm seeing the WP:AFT version 4 up under the track listing section. Until this time, I have seen that particular box at the bottom of the page beneath the various appendices. Using Fx 10.0.2 on Windows 7. I tried shift refreshing the page as well as purging and then refreshing, and it would not disappear. Am I the only one? Or was this a test condition for the box...? --Izno (talk) 19:21, 29 February 2012 (UTC)
Whenever I load a page, it goes completely blank after it finishes loading. Only happens when I'm logged in, works normally when logged out. Did one of you guys break something? If so, can you fix it or suggest a workaround? Kilopi (talk) 23:48, 29 February 2012 (UTC)
Does anyone know how to get rid of all the notations at the top of the page on monobook? ("my talk", "my preferences", etc.). I certainly don't mind them, but this latest revision, like all "improvements", has made Wikipedia even slower and more absurdly difficult to navigate for me. It seems to have to do with these notations, since they're always the ones moving while the page is loading. Is there some kind of code I can enter into Custom CSS and/or Custom JavaScript to get rid of as much as possible without getting read of the text/pictures of every individual article? I already have something for the logo, and getting rid of everything else while keeping article content would be fantastic. All Hallow's Wraith (talk) 05:36, 1 March 2012 (UTC)
The following CSS will remove the top links:
#p-personal{display:none;}
If you want to remove everything, except the "article", "talk", "edit", etc. tabs and the page content, use the following:
Thanks. What about the bottom? On many pages, I get that horizonal scrollbar appearing for a second (or two), and then disppearing, thus taking up more time. How do I get rid of that? (can I?) All Hallow's Wraith (talk) 06:55, 1 March 2012 (UTC)
I'm not quite sure what you mean. The browser's scrollbar? FYI, since you aren't blocking #column-one, you will need only #p-personal { display: none; }. Goodvac (talk) 07:02, 1 March 2012 (UTC)
Well, most people have a vertical scrollbar on the right side of their pages, and then a horizontal one at the bottom (when necessary). "Scrollbar" may not be the right word. Something in the coding on Wikipedia is causing the bottom scrollbar to appear and then disappear every time I open a page (i.e. even when this scrollbar is not necessary, which is most of the time, unlike the vertical scrollbar). I guess the fact that I'm using Internet Explorer '06 is what's causing all this. All Hallow's Wraith (talk) 07:23, 1 March 2012 (UTC)
Well that certainly won't help. I'm predicting that IE6 will officially be dropped in about a year time. It's already down to 3% html requests, when it's below 1%, official support will end. —TheDJ (talk • contribs) 10:21, 1 March 2012 (UTC)
Unable to access scripts
I am unable to edit any page with the scripts I have installed. Whenever I edit a page, the scripts box on the side just remains empty. I have tried blanking and re-creating my scripts page and bypassing my cache, but still I just get an empty scripts box on the side menu. CanuckMy page89 (talk), 06:47, 1 March 2012 (UTC)
Please be advised that document.write is no longer working as of Mediawiki 1.19. Please switch all your scripts and monobooks (if you use vector I don't care about you :P) to the new mw.loader.load. Example:
from
document.write('<script type="text/javascript" src="'
+ 'http://en.wikipedia.org/w/index.php?title=User:Voice of All/adminnolupin/monobook.js'
+ '&action=raw&ctype=text/javascript&dontcountme=s"></script>');
to
<code=>
mw.loader.load('//en.wikipedia.org/w/index.php?title=User:Voice of All/adminnolupin/monobook.js&action=raw&ctype=text/javascript');
If you don't dare change it yourself for fear of breaking it or are not an admin, please reply to this thread and somebody will fix it. Regards, SnowolfHow can I help?07:39, 1 March 2012 (UTC)
Just for your information, the usage of document.write has been ill-advised for over 5 years now. The reason for that was that it would break in async loading. Since we are now loading more parts async then before, this will bring up a lot of issues where people were still using it. This is not a bug. The bug is people after 5 years still going against advice and now finding out. It was not working before, but now more people are being confronted with it not working. —TheDJ (talk • contribs) 09:50, 1 March 2012 (UTC)
Well the problem is not what you had in your monobook.js page. The problem is what the scripts that you were properly loading right there, were doing. One or more of those scripts was/is broken, and the only thing you can do is either find the problem in them (which i guess is not an option), or disabling/removing them until they are fixed by their authors or a kind 3rd person. —TheDJ (talk • contribs) 21:14, 1 March 2012 (UTC)
User contributions dropdown menu
Has thewre been a change in the user contributions dropdown menu? I had this installed, so that I could see user contributions from the top tool bar rather than the side of the page. But this morning, the button had disappeared. I checked my preferences, and installed the "change page and user options" gadget. This reinstalled the buttons; but now, when I click on User:Contributions, I get redirected to Fundraising! Please restore the old functionality. RolandR (talk)08:24, 1 March 2012 (UTC)
I'm not sure about the text, but it's the gadget documented at User:Haza-w/Drop-down_menus. I've now disabled it; but I only added it this morning because the original dropdown buttons and functionality had disappeared overnight. RolandR (talk)09:58, 1 March 2012 (UTC)
The difference between the sizes of the revisions, e.g. (+67) means that that revision is 67 bytes longer than the previous one. ― A. di M.10:38, 1 March 2012 (UTC)
Color doesn't change for links when I visit them; is this an update-related problem?
Moved from WP:RD/C.
When I open a page (in a new tab) in IE 8.0.7601.17514 by middleclicking on an external link from a Wikipedia article, the link typically changes from blue to purple. Suddenly, this color change is not happening: it's just as blue after clicking as it is before. What could be causing this problem? I made some changes earlier today, but several hours after I finished making those changes, the blue/purple issue had not yet arisen. I'm strongly inclined to think that I changed something without meaning to change it. I've not had any updates come through for my Windows 7 system in the last few days. Nyttend (talk) 05:47, 1 March 2012 (UTC)
It sounds like you've switched off browser history, or activated private browsing mode. If the browser is being told not to remember which pages you've viewed, then it won't change the link colour to identify visited links. AJCham08:12, 1 March 2012 (UTC)
I thought that private browsing was only a matter for individual tabs; how does one activate private browsing mode for the entire browser? I forgot to specify that the problem has persisted even after I restarted the computer, so it's definitely not the result of making an error that affects just the specific browser window. Additionally, when I click the tab to bring up the new window, it presents me with a list of closed tabs that I can bring up, just like normal, so I don't think that I've entirely turned off the browser history. Among other things, the "Delete browser history upon exit" button is not checked. Nevertheless, can you tell me how to disable it, so I can disable it and re-enable it? Thanks! Nyttend (talk) 13:01, 1 March 2012 (UTC)
I just remembered something else that might play in. I'm somewhat red-green colorblind, and at least in my case, this means that I simply can't see red or green very well. As a result, I often can't tell the difference between blue and purple, because if the red influence isn't very strong, I simply don't see it at all. Is there a way that I could have told it to reduce the strength of the purple color that will appear when I click the links? I dimly remember from a basic HTML class that it's possible for a website admin to specify the color that will appear when a link is clicked; did MW1.19 contain a change of this sort? Nyttend (talk) 13:23, 1 March 2012 (UTC)
Ugg, it's even more complicated than I realised. At Amos Sawyer, the links to Prince Johnson, Charles Taylor, and Bloomington, Indiana appear in their proper shade of purple; I visited all of them yesterday or the day before. However, even after I visit Ellen Johnson Sirleaf's article, its link is still blue in Sawyer's article. Nyttend (talk) 13:46, 1 March 2012 (UTC)
The reason I've moved this over here is that it may be an upgrade-related problem; I visited a lot of links before 2300 UTC yesterday, and they're displaying fine, while links that I've visited since that time are displaying as if I'd not visited them. Nyttend (talk) 14:18, 1 March 2012 (UTC)
Odd article headers
Hi folks, just wondering if someone could help me out with a small problem, namely that all of my article headers are being displayed with tags around them, in this manner: <span dir="auto">Dextrocardia</span> - what's going on, please? :-S SalopianJames (talk) 14:51, 1 March 2012 (UTC)
Was there a change to the handling of file redlinks introduced in the recent software update to 1.19? If I remember correctly, a click on a file redlink used to go to the edit page of the (missing) file page. While that was in many respects not very useful (we don't want people to edit file description pages of nonexisting files), at least it had one advantage: it showed the user the file's deletion log.
Now suddenly clicking on a file redlink leads to the upload page, Wikipedia:Upload. This is just horrible, for several reasons:
If there is a file redlink, in the huge majority of cases it is because there used to be a file and it was deleted. If a file was deleted, we usually don't want editors to just re-upload it. Even if there was a case for the deletion to be overturned, what we'd want the editor to do is to ask for it to be restored, not to make a new upload.
The upload link doesn't actually go to Special:Upload, but to Wikipedia:Upload. That's the page with the entry to the upload guide. (Right now it's a redirect to the experimental Wikipedia:File Upload Wizard; otherwise it used to be the page that's currently at Wikipedia:Upload/old). The redlink comes with a parameter "?wpDestFile=....", apparently assuming the filename will be preloaded into the upload form. However, unlike the bare Special:Upload, neither the old Wikipedia:Upload nor the current Wikipedia:File Upload Wizard have any use for this parameter. Nothing in the system helps to make sure the re-uploaded file will get the correct filename to turn the redlink into a functioning file link again.
Even worse, neither of the two wizards nor indeed the plain Special:Upload show the user the one, most important piece of information they need to be given: the deletion log. They are thus left entirely in the dark as to why the file is gone.
Thus, in effect, we are inviting people to wildly re-upload files that we just deleted, contrary to our deletion and undeletion policies and processes. We are encouraging them to do that under any new random filename. And we aren't even telling them why we deleted a file in the first place.
Can we please, please get some solution restored that directs these redlinks to something that makes sense? What would make sense is a page that shows the deletion log, possibly with instructions on how to ask for a file to be restored and an overview of our deletion policies or something. Fut.Perf.☼15:30, 1 March 2012 (UTC)
Strange. No, I can't remember ever having been sent to an upload page. It was always either the bare file page address (as in //en.wikipedia.org/w/index.php?title=File:Just_an_example.jpg), or the file page in edit mode (as in https://en.wikipedia.org/w/index.php?title=File:Just_an_example.jpg&action=edit). I seem to remember the former was what you got from a red file-insertion link ([[File:Just an example.jpg]]), the latter from a colon'ed bare text link ([[:File:Just an example.jpg]]). But in any case, the edit-the-file-page solution isn't optimal either, is it? Fut.Perf.☼16:42, 1 March 2012 (UTC)
Ugh. That bugzilla is confusing. So there was something called "$wgUploadMissingFileUrl" which allowed us to do something more useful. What is the effect of this latest change now? I find that hard to read. Is "$wgUploadMissingFileUrl" still around, could we use it, and have we ever used it? Fut.Perf.☼21:49, 1 March 2012 (UTC)
Doubly fixed now. It was a combination of two things:
T36865, resulting in the API returning no interwiki prefixes.
AnomieBOT's code didn't specifically test for "no interwiki prefixes exist" (since that shouldn't ever happen here), and the GIGO result was that it would end up matching the ":" at the beginning of links using the colon trick.
The bug is fixed, and AnomieBOT's code has been adjusted to match nothing if no interwiki links exist. Please don't hesitate to stop the bot again if any other problems come up. Anomie⚔20:28, 1 March 2012 (UTC)
Unable to edit refs
Since the latest changes to MediaWiki, I am unable to edit any refs. If I try to edit a page, instead of the detailed citation, I just see a grey button labelled "ref". There is no way to display or edit the ref itself. Without this, I am totally unable to edit -- please advise and help me! RolandR (talk)20:40, 1 March 2012 (UTC)
Which ones? How? Everything was fine yesterday, so it would appear to be a bug with the latest software upgrade. RolandR (talk)20:57, 1 March 2012 (UTC)
If a user script is non-compatible with the latest software upgrade then the fix will probably be to modify the user script. Try to log out and see if the problems stops. If it does then log in again and isolate the problem. You could for example start by blanking User:RolandR/vector.js. If this makes references work then gradually bring back the scripts until the problem occurs again. Your screenshot is from Beersheba#Antiquity which works fine for me. PrimeHunter (talk) 21:09, 1 March 2012 (UTC)
OK, I've done that, and it seems that WikiEd is the culprit. If I disable it, I can again edit refs. But it's a pain editing without it; what should I do to try to get this fixed? RolandR (talk)21:58, 1 March 2012 (UTC)
You have a lot of scripts installed. As mentioned in the comment immediately preceding your latest one, you should remove all scripts then slowly add them one by one until you find out which one is causing the problem. One of your scripts is causing JavaScript conflicts. You could also check Firefox's Error Console to see if there are any errors that might give you clues as to what script is causing the problem. Gary King(talk · scripts)22:01, 1 March 2012 (UTC)
I've done that, and established that WikiEd is causing the problem. Currently reporting it on the WikiEd bug report page. RolandR (talk)22:16, 1 March 2012 (UTC)
And I've received an answer. My WikiEd settings had somehow been changed (presumably by the upgrade, as I changed nothing overnight), and I've now restored full functionality. RolandR (talk)22:58, 1 March 2012 (UTC)
ESC cancels out all edits
I'll keep this brief since I just lost my last message. In IE8, ESC cancels all wiki edits. I just lost a lot of work after I hit something that popped up a window. I hit ESC one too many times, and I lost all my edits. This is not an issue on Firefox, but I sometimes edit on a computer that cannot have anything besides the approved software bundle. Any workarounds besides save early, save often or working on a separate text editor? Cheers. Encycloshave (talk) 21:25, 1 March 2012 (UTC)
Yeah, I've noticed this several times before. I don't have any suggestions (beyond what was given above) except to be very careful when you're signing your posts or typing the "1" key. I do almost all of my writing in a text editor, as you mentioned above. Nyttend (talk) 03:16, 2 March 2012 (UTC)
Many thanks Alex!
popups and Twinkle have disappeared ...
... in Firefox (my preferred browser).
Twinkle has also disappeared in Opera, but popups works fine there.
Both are working fine in IE9 and Chrome.
Using Monobook skin (appears to be the same for Vector, but haven't checked all possibilities yet).
Thanks, that fixed it (Sometimes it pays to check the obvious! I gave up this morning, when I saw what a total disaster Wikipedia was ...). --NSH001 (talk) 22:25, 1 March 2012 (UTC)
How to regenerate an image thumbnail?
For the thumbnail here, the bottom third or so is cut-off; it just appears white. How can I purge the thumbnail so that it is regenerated? I tried purging the page, then the image (on Commons), but neither method seems to do the trick. Gary King(talk · scripts)03:34, 2 March 2012 (UTC)
All indicated images look fine to me, suggesting that it's a browser issue than WP's. Trying hard-reloading the page (usually shift-F5 for most browsers). --MASEM (t) 03:47, 2 March 2012 (UTC)
When the thumbnails were recently moved to the new swift architecture, some got corrupted in this way. Apparently the bug causing that has been fixed, but some bad cached images remain. Purging them should fix it, if it doesn't, then that means purge signals are again not reaching one of the caching servers. —TheDJ (talk • contribs) 07:15, 2 March 2012 (UTC)
Sounds right. I've purged the Black Sox several times, and full purging from the browser seems to have finally done it. The Hobart image has cleared, too; not anything I did, though. Thanks, Alarbus (talk) 07:40, 2 March 2012 (UTC)
Hgrosser's been wrapping iw links with a span styled display:none; He's also tried baking-in {{link FA}} IDs: Diff of Murder of Julia Martha Thomas. This seems in good faith, but I don't think it a good idea. He's done this on a bunch of wikis; fa, pt, zh, hi, he, da, de...
I only noticed this because he did it to the current main paged article. This is going to need more participants on that help page or a better venue. Alarbus (talk) 04:20, 2 March 2012 (UTC)
The developers probably saw no reason for them to have subpages. They are files, after all, not pages of content. This subsequently allows files to have slashes in their names if necessary; I don't know exactly how common that is, though. Gary King(talk · scripts)01:18, 2 March 2012 (UTC)
Slashes cannot be part of a filename because they separate directories from filenames. So, if an image were named File:Foo/bar.jpg this would actually be a file named bar.jpg residing in a directory named Foo. You can see how Mediawiki uses directories to manage files by right-clicking on any exposed image and selecting "Copy Image Location". Then paste that into any plain text editor (such as a Wikipedia edit window). You might get something like this: http://upload.wikimedia.org/wikipedia/commons/thumb/5/5b/Kate_Webster_filtered.jpg/77px-Kate_Webster_filtered.jpg - here, 77px-Kate_Webster_filtered.jpg is the filename, http://upload.wikimedia.org/ is the base URL, and wikipedia/commons/thumb/5/5b/Kate_Webster_filtered.jpg/ is a chain of six directories, the path through the directory structure. --Redrose64 (talk) 17:33, 2 March 2012 (UTC)
Removing a page from the search box
Is it possible to remove a page from the search box? For example, when you type something into the search box several options are shown to the page that you may be looking for, is it possible for a page to be 'invisible' per se? Whenaxistalk·contribs | DR goes to Wikimania!02:24, 2 March 2012 (UTC)
There's the behavior switch "magic word" __NOINDEX__, intended to prevent search engines from indexing a page. But I don't know whether it affects Wiki's own search results. — Richardguk (talk) 04:31, 2 March 2012 (UTC)
How Are Non-Logged-In Users Able to Make Any Edits
How in the world can users that aren't logged in able to make any edits? When I'm not logged in, I don't see any way to make any kind of edits (read "View source"). Thanks! Allen (talk) 03:13, 2 March 2012 (UTC)
You are only supposed to see "View source" on the relatively few protected pages. Here is a random unprotected article: Jashwant Singh. I see "Edit" when I log out. There once was a bug where unregistered users sometimes saw "View source" on unprotected pages. They could click it and edit the page anyway, but I haven't seen reports of this bug in a long time. PrimeHunter (talk) 03:20, 2 March 2012 (UTC)
The redirect button appears to have disappeard from the list of buttons above the edit window. Can we have it back please? Mjroots (talk) 05:15, 2 March 2012 (UTC)
Click Advanced and then the right-pointing arrow. It's also among the options when you select Wiki markup in the drop down box below Save page. PrimeHunter (talk) 12:38, 2 March 2012 (UTC)
I use the old toolbar, and the redirect button indeed disappeared after the upgrade to 1.19. But just a few hours ago, it reappeared. Goodvac (talk) 05:52, 3 March 2012 (UTC)
When I've got the edit window open, there is no "advanced" feature that I can see. Where should I be looking. My browser is Firefox, BTW. I've not changed any settings, so I don't know which toolbar I'm using. *sigh* If it ain't broken... Mjroots (talk) 06:58, 8 March 2012 (UTC)
Then you must be using the old toolbar, which looks like this. When I enter the editing mode, the redirect button does not appear; when I hard-reload the page, then it appears. To see the redirect button, I have to hard-reload every time I go to an edit window, so there is something broken, but I don't know what. Goodvac (talk) 07:44, 8 March 2012 (UTC)
Seems to be related to #Missing toolbar buttons? further up among the MW 1.19 bug list. Anyway, the straight row of 23 buttons is RefToolbar 1.0, which doesn't have the "Advanced" thing. That is in refToolbar 2.0: you can see which you have, and see how to switch between them, by checking these features. --Redrose64 (talk) 12:45, 8 March 2012 (UTC)
I've now turned on the 2.0 version, and got the redirect button back. That doesn't mean that the original issue raised has been fixed though, Mjroots (talk) 17:18, 8 March 2012 (UTC)
Currently hitting the delete key on the image does not return anything for over 20-30 seconds, often resulting in "wikimedia foundation error..." The tech details are oft similar to "Request: POST http://en.wikipedia.org/w/index.php?title=File:Logo_FEPLP.jpg&action=delete, from 208.80.154.8 via cp1007.eqiad.wmnet (squid/2.7.STABLE9) to 10.64.0.140 (10.64.0.140)
Error: ERR_READ_TIMEOUT, errno [No Error] at Fri, 02 Mar 2012 06:09:30 GMT". The 208.80 is familiar, but the 10.64 isn't (Durham, NC?). A ping from me to 208.80 is consistently 99ms, so it doesn't seem to be a timeout from this end. Skier Dude (talk) 06:17, 2 March 2012 (UTC)
As I said (just a moment ago, after you said this), I'm trying to change it for myself because it seems that the default for clicked and unclicked has now become the same color. See the "Color doesn't change for links when I visit them; is this an update-related problem?" section above, which I posted recently; I went to the Help Desk because that other thread has been completely ignored here. Skin is Monobook. Nyttend (talk) 13:43, 2 March 2012 (UTC)
I'm using Monobook, but the default CSS already has .mw-body a.external:visited{color:#636; }, which should take care of it. The problem is it isn't working, even if I add the same code to my monobook.css. Goodvac (talk) 18:50, 2 March 2012 (UTC)
Actually, maybe the problem is specific to Monobook. If I go on Meta, where I use Monobook, clicked external links do not change color. On other wikis, such as French, where I have the default Vector, the links do change. Goodvac (talk) 18:54, 2 March 2012 (UTC)
I am using the ‘Chick’ skin, and with this skin every Wikipedia page used to have two links ‘Jump to: navigation, search’ at the top. I don't see these links anymore, and I cannot make them to appear even by explicitly setting the option ‘Enable "jump to" accessibility links’ in my preferences. (For a description of this option see meta:Help:Preferences#Advanced_options.) I have tried this with Firefox 10.0.2 and Internet Explorer 9. I have cleared the browser cache, with no change. Is this something that was removed with the recent rollout of MediaWiki 1.19? — Tobias Bergemann (talk) 08:54, 2 March 2012 (UTC)
I'm after a list of our infoboxes, sorted by number of transclusions; specifically the low end, so I'd settle for a list of those with, say, fewer than 250 instances. Do I need to ask on BOTREQ, or find someone with toolserver account, or is there a better method? Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits17:31, 2 March 2012 (UTC)
That sounds like a Toolserver request. If you drop a note on my talk page, I can probably help you out. It's a fairly simple query. --MZMcBride (talk) 18:28, 2 March 2012 (UTC)
I was trying to leaving a welcome message on User talk:Kerev HaEmet via Twinkle. After a fair amount of time, it ended up at the techical error page. Now, every edit I try on the talk page is resulting in an edit conflict (including adding a new section), yet there's nothing in the history for the page. I suspect something's messed up (database?), but it's beyond me. Any ideas on this? Ravensfire (talk) 19:13, 2 March 2012 (UTC)
I deleted the page, so it should be editable now. It must have been some kind of glitch, since there was no history for the page. Reaper Eternal (talk) 19:20, 2 March 2012 (UTC)
There was some downtime a few minutes ago. I have a slight suspicion that the create+edit trough the API as issued by Twinkle was not atomic, causing the api to do the create, but not the edit and then not rolling back. That would be a bug in the API of MediaWiki. —TheDJ (talk • contribs) 19:24, 2 March 2012 (UTC)
Interesting ... when I'm mentoring new developers at work, transactional boundaries are one of the first things I hammer into them as far too many schools skip on that and it's really easy to miss something. (Especially with an older app that's still using EJB 2.1 ...) Dealing with boundaries with web services / api's can be a nightmare though - I've spent enough of the past 2 months dealing with those (rolled out the first major service at my company) to feel some empathy. Ravensfire (talk) 19:29, 2 March 2012 (UTC)
Heavy load time for deleting files
For the past month or so, Mediawiki has been so slow in deleting files. What happens is the file will be deleted immediately, but the confirmation isn't returned form the server for a very long time. Sometimes it just takes a long time and it works; sometimes it will fail altogether. This is a huge pain for me because I do Twinkle deletes of NowCommons files by the hundred, and it can take a half hour or longer now for a set of images.
Asking around IRC, it appears the issue doesn't exist with regular pages, just files. Anyone with administrator privileges can verify this by deleting an image in CAT:NCT.
Example error code for reference below.
Request: POST http://en.wikipedia.org/w/index.php?title=File:Matchtosample.png&action=delete, from *REDACTED* via cp1005.eqiad.wmnet (squid/2.7.STABLE9) to 10.64.0.138 (10.64.0.138)
Error: ERR_READ_TIMEOUT, errno [No Error] at Fri, 02 Mar 2012 20:27:48 GMT
This is a strange problem I've just been having now (on Monobook). Some parts of the page fail to load completely. The top bar is on the side, the big globe doesn't show up, and the page looks strangely stripped. This used to just happen in the article space (Main Page and VPT were OK), but nwo it's happening on the "new section" VPT page. hbdragon88 (talk) 05:29, 3 March 2012 (UTC)
Hmm, nothings happening with me in monobook. Have you tried removing all the scripts in your script files and adding them back in one-by-one to see if it is one (or the combination) of them causing the problem, as script files have caused several things not to display properly for me in the past.--GilderienTalk|Contribs22:02, 3 March 2012 (UTC)
Templates for Displaying Military Ribbons/Medals as Worn?
Are there any templates for displaying a user's military ribbons and medals as worn on his uniform? If not, how can I do it? User:Gadget850 said that he uses the template "Quote box" to place ribbons into. Is that good, or is there a better way? Thank you. Allen (talk) 17:27, 3 March 2012 (UTC)
Aggregate pageviews for 10 different articles
I'm trying to compile a small dataset of the total pageviews articles I've created have received. I know I can use stats.grok.se, but it only pulls individual months (or 60/90 day periods) at a time. Is there a way to get more comprehensive data that doesn't involve querying the entire database? I'm not a coder. Thanks! Ocaasit | c17:34, 3 March 2012 (UTC)
Look at the URL for a monthly stats page — you can see both the year and the month number. Simply delete the month number from the URL and it will give you stats for the whole year, as long as the servers were working without hiccups. Nyttend backup (talk) 04:06, 4 March 2012 (UTC)
Thanks for the tips, folks. The 'removing the month' from the url trick doesn't work; it returns a 404 error. Jarry, I'm checking out your script. I think it could save me a lot of time. Ocaasit | c13:48, 4 March 2012 (UTC)
Interlanguage links
fr:Personne physique seems to have correct interlanguage links, but the linked articles (including English and German until I fixed them) link back to fr:Personne morale. These wrong links seemed to have been inserted by a bot. Is there a way to get a bot to fix all of them instead of making them wrong? —teb728tc21:37, 3 March 2012 (UTC)
Thanks for that. Based on that the change I made might work by itself. However, the toolserver account for the linked tool has expired. —teb728tc10:22, 4 March 2012 (UTC)
Disabling Javascript
Is there code I can enter into User:All Hallow's Wraith/monobook.css in order to disable Javascript on Wikipedia? (without having to to so for my whole computer). I have a sinking feeling Javascript is behind why Wikipedia turned into a soul-suckingly slow monstrosity around roughly February 2011. All Hallow's Wraith (talk) 03:07, 4 March 2012 (UTC)
CSS won't be able to do that. If you use Firefox, however, you can install YesScript or change your prefs.js (harder). There are probably extensions that can accomplish what you want for other browsers as well. Goodvac (talk) 03:27, 4 March 2012 (UTC)
Opera has a toggle for turning on and off js in its quick preferences menu, and chrome was, last I checked, less annoying than firefox when it comes to its preferences (probably has addons too, but I dunno about that stuff), if you use either of those or are shopping around for browsers. — Isarra (talk)08:37, 5 March 2012 (UTC)
I'm an admin and don't see entries like that in the last 5000 entries at Special:AbuseLog. If I log out then I see it for some entries, for example the most recent at [10] but not the 4 others. The hidden entry is for a filter which enables "Hide details of this filter from public view". This feature is for example enabled in Special:AbuseFilter/384 and disabled in Special:AbuseFilter/432. I don't know whether non-admins used to be able see more in log entries for non-public filters. Wikipedia:Edit filter has said for a long time: "Log entries are viewable by all users, and while filters are by default publicly viewable, others are set to be private. For all filters, including those hidden from public view, a brief, general summary of what the rule targets will be available, and displayed in the log, the list of active filters, and in any error messages generated by the filter." PrimeHunter (talk) 21:58, 4 March 2012 (UTC)
The problem is that I can't see filter triggering; should I just apply to be an edit filter manager to be able to view those entries?Jasper Deng(talk)02:03, 5 March 2012 (UTC)
Yes, it was one of the changes in MediaWiki 1.19, as noted in the release notes under the section Configuration changes in 1.19. Looks like the reason for the change was Bumped $wgCookieExpiration to 180 days, we were one of the sites with shortest cookie lifetime for too long. This was discussed at lengths about a year ago on wikitech-l, however nobody cared to implement the suggested alternative solutions. Probably because they were overcomplicated and solved only parts of the problem[11] - Kingpin13 (talk) 08:24, 5 March 2012 (UTC)
I've created my User:Sameboat/vector.css to change the look of the wikitable class table. But no matter how many times I purge and refresh, the look of the wikitable just won't change. Is there something wrong with my css? I'm using Vector skin really. -- Sameboat - 同舟 (talk) 10:26, 5 March 2012 (UTC)
(edit conflict) I see you're trying to change the color of the borders. The rule for changing the border of the <table> element is correct, but has no effect since the borders are set to 'collapse' which causes the border colors set on the individual cells override it. The rule changing the color of the border for the cells is not working because the selector for the cells, table.wikitable td, table.wikitable th, has a lower "specificity" than the rule in the site CSS, table.wikitable > tr > th, table.wikitable > tr > td, table.wikitable > * > tr > th, table.wikitable > * > tr > td. Your best bet would be to copy that selector. Anomie⚔12:25, 5 March 2012 (UTC)
!important is also something to be avoided however; it is a last resort. The proper way is to use the proper selector to target the part of the layout that needs to be targeted. For instance, your definition might also override certain inline CSS styling in articles which you would like to see and normally expect to be applied, even with your changes. It will target more parts of the layout than wikitable would otherwise target. The reason the selector of wikitable is so specific is exactly to avoid such unwanted behavior. just so you know.... —TheDJ (talk • contribs) 00:00, 6 March 2012 (UTC)
I reported this several months ago, but do not recall receiving a response. After making an edit, saving it, and then removing a category using HotCat, the previous edit I made is reversed, as can be seen in this diff. Does anyone have any idea why this happens, and, more importantly, how it can be corrected? Thanks! ---RepublicanJacobiteTheFortyFive19:21, 5 March 2012 (UTC)
Invitation to help test the new MediaWiki extension for the Education Program
Our developers have been working hard to integrate certain elements of the Wikipedia Education Program into MediaWiki. If anyone is interested in helping test the new extension, click here to get started.
Fully agree - when on an archive page and wanting to check back to what the article currently says it leads to an article creation page for Subject:Archive. I cannot envisage any instance where we would want an article page solely relating to the archive - archive pages should link back to the actual article to facilitate checking archives against the current version. - Arjayay (talk) 21:57, 5 March 2012 (UTC)
Technically, there currently is no such thing as an archive page. Technically, there is only a talk page that functions as an archive of discussion that took place on another talk page, which might or might not be bound to a specific "non-talk" page. Which is why you observe the behavior you describe. Options; Status quo, Javascript hack, Liquid threads or changes into the core software (of which most developers will probably say: "better wait for LQT-nextgen"). —TheDJ (talk • contribs) 23:49, 5 March 2012 (UTC)
How about just using a bot to create a "nonprintworthy" redirect to the actual article, for each missing mainspace page that corresponds to a talkspace subpage? — Richardguk (talk) 02:19, 6 March 2012 (UTC)
And where does the link go when you're on Talk:Article/sandbox? This isn't something that needs to be fixed, because it's not broken—just follow the breadcrumb links back to the main talk page. — Bility (talk) 16:31, 6 March 2012 (UTC)
I think it is. A non-expert user clicking the button will be left in a blank editor, with no "breadcrumbs". Only those with some a posteriori knowledge of the layout will be able to know what to do. After all, it confused me, and that's saying something. Maury Markowitz (talk) 17:10, 6 March 2012 (UTC)
If we really wanted to deploy a JS hack, it would have to be a bit smarter that this. I'd propose loading on all talk pages. I'd add a recognizable classname to all archive page headers and then conditionally upon that change the link. The below code is untested, and not working, but it's an adequate representation of the idea I have. There are probably some JS functions in the new RL core to better deal with these page titles, and the talk -> nontalk page switch. —TheDJ (talk • contribs) 23:27, 6 March 2012 (UTC)
Or the user will hit the back button in their browser, follow the breadcrumbs and realize how talk pages work. If it's a gadget or userscript that's fine with me, I just wouldn't personally want to add to Wikipedia's already bloated JS. — Bility (talk) 00:04, 7 March 2012 (UTC)
A very strange problems has arose for me on Wikipedia. On every page the title displays as Page Title, where page title is the title of that page. It has been going on for about a week and a half now. I didn't bother coming here until now, because it hadn't annoyed me enough. This problem occurs on every page, including my watchlist, preferences, and contributions pages. I checked on article pages and there is no span tags around any titles at the top of the page or the bottom. Don't know how to fix it or what is going on, and looking for a little help.--NavyBlue8414:01, 6 March 2012 (UTC)
There is a problem with an embedded blank showing up where there should be none.
Look to the 2nd paragraph beyond the one you are reading, then back up one
paragraph to learn how I arrived there.
I didn't think I should use the "Talk -- Main Page" for this, but there is no
place I can comment about a format problem turning up when I go to a "Special
Page". When I enter "groer" (no quotation marks) in the search prompt on the main
page, I get taken to a special page which says there is no page called "groer"
(that's OK), and provides some possible entries, including the one about the
now-deceased Cardinal from Austria which is included below (at bottom of message
you are reading) with some line breaks lost when I copied and pasted.
THE REASON I AM WRITING is that I have been noticing an embedded blank (in the
example included here, it turned up in the word "Austrian") on such Special Pages.
Why is that embedded blank showing up? Here is the output I am asking about:
Hans Hermann Groër
Hans Hermann Wilhelm Groër, OSB (1919-2003) was an Austria n Cardinal of the
Roman Catholic Church . He served as Archbishop of Vienna ... — Preceding unsigned comment added by 128.63.16.82 (talk) 17:15, 6 March 2012 (UTC)
It's not a typo. It's related to the way wiki links can be made. If you look at the wikitext of the article in question you'll notice that the relevant bit of text is [[Austria]]n which displays in a page fine like so: Austrian. It would appear that search strips the square brackets and adds a space. As such I suspect this is a minor bug in search. Dpmuk (talk) 18:17, 6 March 2012 (UTC)
Suffix gaps in Search are a feature not a bug: Wait a minute, I like seeing Wikisearch results show "Austria n" for "[[Austria]]n", and an editor could write "[[Austria|Austrian]]" to avoid the suffix gap at "n". One's man's bug is another man's special feature. I worry that changes to Wikisearch will also hide "abbr=on" in searches to detect "{{convert|2|km|mi|abbr=on}}" where a new "bugfix" would prevent searching for "abbr=on" in that manner. Be sure to consider a new option such as "Classic Wikisearch" to retain the old "features" because too many things have gone wrong in efforts to drop support for old browsers, or jump-start new gadgets that few editors use. I wish Wikipedia still worked like it did back in January 2011, and BTW, I do not need autocompletion of words entered for Wikisearch. However, if the developers are bored, I would like them to write a WP:Parser function to insert a real &minus sign in "-50.0400" or "-5,000,200,000,000" and retain the trailing zeroes and not reformat as scientific notation, -5.0002E+12 and such. Also, everyone knows we need "{{set:xk|45}}" to set a parameter "xk" to have value "45" during the evaluation of a transcluded template. -Wikid77 (talk) 23:55, 6 March 2012 (UTC)
At this writing, I have gone back to the article itself (about Hans Hermann Gro:er), where I see "Austrian", with the cursor showing that it contains a pointer to "Austria" article. When I go into edit mode in that article (with no plans to change anything), I am seeing "Austrian". — Preceding unsigned comment added by 128.63.16.82 (talk) 16:41, 8 March 2012 (UTC)
Hi. Is there a way to have collapsible sections without using the navbar? i.e. be able to collapse the text underneath == Heading 1 == so only the heading is visible unless you click the [Show] link. Thanks, JhfireboyTalk01:29, 7 March 2012 (UTC)
Please assist There are a variety of English variant templates (e.g. {{British English editnotice}}, which links most, if not all of them at the bottom) and they are themselves composed of templates within templates and transcluded documentation pages. The problem is that the doc pages produce text of the sort (e.g.) "Placing this template will also add the page to Category:Articles which use British English", when in reality, the proper categories are of the sort Category:Wikipedia articles that use British English. I can't figure out the byzantine architecture of these templates to fix this. Help? —Justin (koavf)❤T☮C☺M☯ 07:31, 7 March 2012 (UTC)
Hi. I can't seem to get the references working on this template. I have set up an #if decision where is there is no reference written then it won't show the reference at all. That works, but the reference is blank when you view it on an article. Many thanks, JhfireboyTalk20:31, 7 March 2012 (UTC)
Fixed You can't use variables in parser tags. You can use #tag:ref, or more simply {{refn}} which implements it more easily. You really need to use a group name, else you risk mixing the table references with the main references. ---— Gadget850 (Ed)talk20:45, 7 March 2012 (UTC)
How many of these weird little "I can't figure out <ref> or {{#tag:ref}}" templates do we have now? {{r}}, {{refn}}, {{sfn}}, {{sfnm}}, {{sfnp}}, {{efn}} etc.? And they all break scripts and bots that deal with refs in wikitext, and are liable to confuse editors who have managed to figure out that references are in <ref> but aren't familiar with these myriad templates. Anomie⚔23:26, 7 March 2012 (UTC)
I've noticed that in Wikipedia search results, most wikitext markup is stripped in the text that provides context for the search hit. For example, if I'm searching for "Dog", on the entry for Dog in the search results I'll see something like "The domestic dog, Canis lupus familiaris is a subspecies of the gray wolf" in the search results, but the wiki text says something like "The '''domestic dog''', '''''Canis lupus familiaris''''',<ref>{{Cite|...}}...</ref> is a subspecies of the [[gray wolf]]". How do all those templates, tags, brackets for links, etc. get removed? When I search on many other MediaWiki wikis, I tend to see a lot of that markup in the search results, but in Wikipedia most of the markup is removed (I occasionally see some template parameters, but that's about it). Is there an extension or hook or something else that cleans up the search results? Or is there a way to configure MWSearch to return cleaner results? --Llarq (talk) 22:06, 7 March 2012 (UTC)
Can anybody work out why the bot keeps listing Hexxen, Hexen, and File:Star.jpg as being broken redirects (redirects to red links), despite all three appearing to be perfectly functional? It's been doing it for a few days, so I suppose it's not impossible that it's related to the 1.19 rollout. Thanks, HJ Mitchell | Penny for your thoughts? 02:11, 8 March 2012 (UTC)
Until recently I used to have an old 4:3 aspect ratio laptop. I permanently ran IE in a maximised window; and I think in about ten years I needed to resize it about three times. Now I have one of these 16:9 ratio screens, which are stupidly unergonomic for anything other than watching videos, and I seem to spend half my life fiddling around with window size and zoom adjustments just to get something I can read comfortably. Some websites seem to need a big width to avoid the need to scroll horizontally (which is horrible and to be avoided), and yet when a Wikipedia page is displayed in the same window, the lines are ridiculously long and very hard to read. Which brings me to my point about Wikipedia. Given the ubiquity of these stupid 16:9 screens, should we have a preset maximum width of a Wikipedia page, so that you can maximise the browser window and have an article displayed, perhaps centred, at a reasonable size, with suitably-sized blank margins either side? [12] is an example of what I mean. IMO this is a better idea than letting line lengths extend indefinitely. 86.146.109.211 (talk) 00:01, 9 March 2012 (UTC)
Search for extension?
I want to clean up some timelines, but I need to learn how to use the legend attribute. I thought it would be easiest to look at some existing examples. I tried searching for <timeline>, but search will look for the rendered text, not the code. Is there a way to search for articles using the WP:TIMELINE extension?--SPhilbrick(Talk)13:02, 9 March 2012 (UTC)
Does anyone know (or know how to find out) what proportion of users have access to Unicode fonts built into their operating system? Such users would, for example, be able to see this: ფირცხალაიშვილი as Georgian script without doing anything special. Wikimedia has access statistics for 2011, including breakdown by operating system, but I don't know which of those will be able to display the script, so I can't produce an overall figure. Thanks, Rd232talk21:25, 9 March 2012 (UTC)
I am not an expert, but isn't "access to Unicode fonts" too vague to be useful? I have "access to Unicode fonts" in the sense that most scripts (including the Georgian above) display OK, but I still sometimes see square boxes because the fonts do not have that particular character. In fact, I think few fonts would support every single character defined by Unicode. 86.146.109.211 (talk) 23:40, 9 March 2012 (UTC)
There are varying combinations too. For example, Chinese, Japanese and several other Asian fonts render just fine for me; but a frequent contributor to this page has ⚔ as their talk page link, and for me that's a square containing the numbers 25 and 94. My mother's laptop, on the other hand, cannot produce Chinese or Japanese script, but renders Anomie's talk page link just as it's supposed to look. --Redrose64 (talk) 14:10, 10 March 2012 (UTC)
I see. Well, the background to this is that there is an issue on Commons on whether there are circumstances (if so, what) where text in image form can be kept because it's useful for some people. Generally, it's assumed that people can see (most?) Unicode text, and that therefore text in image form is completely unnecessary, and it's subject to deletion. I'm just trying to get a bit more information about how widely that assumption holds. Rd232talk16:27, 10 March 2012 (UTC)
New Page Triage
Hey guys :). As previously mentioned in a few places, the Foundation has started work on our new patrolling software: New Page Triage. I'm posting updated specifications in a few hours, and I'd really advise everyone who is interested in page patrolling to head over to the talkpage, comment on the suggestions on the page already and the additional ideas the community has come up with.
We've also got an office hours session next Tuesday, the 13th, at 19:00 UTC (that's 12:00 PST, for the west-coast Americans around ;p). If you can make it, it's on IRC in #wikimedia-office. If you can't, drop me a line on my talkpage and I'm happy to send you the logs once we're done :). Regards, Okeyes (WMF) (talk) 21:43, 9 March 2012 (UTC)
It used to say "This user assumes good faith" but then I decided to add "inside and outside of Wikipedia". I worked, but the wikicode got all messed up. I'm hoping I'll get a message about someone reverting it. Any help?
I've reverted it for you. In general, you shouldn't really change existing userboxes that are already used by lots of people, as you might be changing them to something unexpected that the people using them do not want. If you want a variation on a userbox, it's better to create an entirely new one. -- Boing! said Zebedee (talk) 13:51, 10 March 2012 (UTC)
I talked on the talk page a bit, and I noticed I pointed out this to another editor there. The conversation can be seen here. He saw my point, and if there are third-party reliable sources, then that age me the green light to go ahead and divide the episodes into seasons. So I made myself a preview, but it didn't work smoothly. If you clicked the link that said "here" you would even see I noted I couldn't work up the wikicode to do it. Any help?
The NewDiff gadget has been updated and is now based on r133098, designed by Trevor Parscal, which is currently in trunk, with some minor adjustments. Comments welcome. — Edokter (talk) — 14:54, 10 March 2012 (UTC)
Yes, it should. In {{trivia}}, the code {{DMCA|Articles with trivia sections|from|{{{date|}}}}} should be changed to {{DMCA|Articles with trivia sections|from|{{{date|}}}|All articles with trivia sections}}. The bolded part is parameter 4 of {{DMCA}}. Goodvac (talk) 23:24, 1 March 2012 (UTC)
Yes I reverted Drilnoth's bold adding of the category. We have enough of these categories as it is, they clutter the hidden category list, see picture. RichFarmbrough, 09:38, 2 March 2012 (UTC).
...in some extreme cases, but most articles have only two or four of these (and they are hidden for a reason). Removing some of the excessive tagging from such pages is more helpful than removing the cat from all of them. Fram (talk) 12:04, 2 March 2012 (UTC)
OK, do you patrol problematic categories? Do you even have the display of hidden categories turned on? Working with these categories day in, day out, the multiplication of categories is a problem, and the reasons for having "all" categories are largely or entirely obsolete. RichFarmbrough, 13:52, 2 March 2012 (UTC).
Yes I do, and yes I have. That you don't have a use for these "all" categories doesn't mean that they are not needed or useful for other people. Fram (talk) 14:46, 2 March 2012 (UTC)
I've undone the removal of the "all" category, until there is actual consensus that it isn't wanted. Fram (talk) 07:58, 2 March 2012 (UTC)
Likewise I undid the addition of the "all" category until there was consensus that it was wanted. It's supposed to be BRD, not BRRD. RichFarmbrough, 13:52, 2 March 2012 (UTC).
You can presume what you like, and doubtless will. The reasons you gave for keeping them then are as wrong now (or more so). RichFarmbrough, 20:09, 3 March 2012 (UTC).
These still don't add up - 102 pages in the "all" category, but the by-months add up to 173. I've tried purging the cache and the numbers stay the same. Nikkimaria (talk) 22:54, 11 March 2012 (UTC)
I looked at the Top Ten list at Wikipedia:Reader (source [13]) which is bizarrely topped by two rather obscure people, Robert L. Bradley, Jr. and William Kurtz Wimsatt, in a "Special:Export" mode. The pages themselves are viewed 337 times/30 days and 6 times/30 days. Was this some kind of DDOS attack with a random target, or can the export feature be otherwise abused by specific people? Wnt (talk) 20:35, 8 March 2012 (UTC)
I don't know what happened there. Not that this addresses what those guys are doing on the list at all, but from the content it looks like a December 2010 list, rather than May 2011 as Wikipedia:Reader says. --SubSeven (talk) 22:19, 10 March 2012 (UTC)
I use Helferlein's macro regularly. (See first entry in external links). It works fine, but if someone has better options, I'm listening.--SPhilbrick(Talk)20:09, 10 March 2012 (UTC)
I have an OpenOffice.org Calc spreadsheet, and I want to convert it into a table that I can put on Wikipedia. However, I want to preserve as much of the formatting (font family, font size, text color, etc.) as possible. If not, I would like to know the wiki code for the formatting. If you know how to do what I would like, I would appreciate it. Allen (talk) 20:52, 10 March 2012 (UTC)
Weird thumbnail errors
at Old Finch Avenue Bailey Bridge, the image thumbnail in the infobox seems to be only half loading for me and several others. The full size image loads fine, as does the File: page... It's just the thumbnail. A similar problem occurs for me at Ontario Highway 420, where the image on the right directly below the infobox appears slanted. I uploaded a rotated correrction to this image several months ago, but the image won't update. Any insights? - ʄɭoʏɗiaɲτ¢18:06, 10 March 2012 (UTC)
Adding a confirmation dialog to the watchlist rollback link
One too many times I've accidently misclicked a rollback link in my watchlist, either on my narrow mobile phone which makes each entry on my watchlist occupy several lines and can lead to misclicks, or because I am meaning to hit one of the links immediately next to it. Stopping your browser after clicking the link makes no difference if you have any decent-speed internet service. When viewing a diff, I believe twinkle or some other add-on introduces three rollback links (AGF and vandalism are two of these rollback links), and these all bring up a dialog box that makes you confirm that you want to rollback and allow you to enter a reason in one instance. Can this be added to the watchlist link to prevent misclicks? - ʄɭoʏɗiaɲτ¢02:06, 11 March 2012 (UTC)
Yet again, the search index has failed to update for at least three days. Help:Searching#Delay_in_updating_the_search_index says this should be reported here. These backlogs frustrate us WikiGnomes in our tidying up, and we get a huge backlog of spelling, grammar and other mistakes to correct when it is eventually updated. Any idea when it might be updated? Arjayay (talk) 18:58, 11 March 2012 (UTC)
Is there any way of search Wikipedias references/citations for particular authors, journals, publications etc. as references seem currently to be excluded from the "everything" search option.--GilderienTalk|Contribs21:54, 11 March 2012 (UTC)
References are written in the main namespace which is searched by default, but if you put quotes around an author name then you may miss citation templates which separate the first and last name in the page source. I don't know a way to search only the references in the mainspace and not the normal article text. "Everything" means all namespaces so it will give you more "false" hits than the default search. PrimeHunter (talk) 00:19, 12 March 2012 (UTC)
[www.w3schools.com/css/css_font.asp CSS Font], at w3schools. Apply as "style="font: foo"" to the table element of the table. --Izno (talk) 00:05, 12 March 2012 (UTC)
You have already received advice at Wikipedia:Help desk#Changing font. If you go ahead and add a mainspace table with various fonts for no other reason than to make the text look different from normal articles then other editors may remove the fonts. PrimeHunter (talk) 00:12, 12 March 2012 (UTC)
First, what does the "www.w3schools.com" thing do?
Second, yes, I have asked this on the Help desk, but they haven't really given me any decent advice.
Third, you misunderstood me. The table is not in the mainspace. It is one of my user pages (User:Morriswa/Highways). The reason I use different fonts is to more easily identify each kind of highway in the list. I could upload a copy of it so you could see what I mean. Allen (talk) 00:49, 12 March 2012 (UTC)
I'm still no closer to getting this done. I'm trying to go through the table manually, but that will take forever. Also, I don't know how to format some things, such as changing fonts. Allen (talk) 01:37, 12 March 2012 (UTC)
Are you planning to move or copy User:Morriswa/Highways to a mainspace article? If no then why do you care so much about the fonts? If yes then have you read the guidelines I linked which are against changing the font? Here is the deprecated <font>: The quick brown fox jumps over the lazy dog. 0123456789. Here is code going against Wikipedia:Manual of Style/Text formatting#Font family: The quick brown fox jumps over the lazy dog. 0123456789, The quick brown fox jumps over the lazy dog. 0123456789. PrimeHunter (talk) 02:19, 12 March 2012 (UTC)
Is there any way that http://stats.grok.se/en/latest/Cuisinart can be accurate? Cuisinart was only edited four times last year, it has seven sentences, 1.7 KB, 2 refs, 2 ELs, a photo, and a navbox. How can it possibly get a million hits per month?
I would not be surprised if it turns out some program somewhere is using that article to test network connectivity or something along those lines. Anomie⚔00:38, 13 March 2012 (UTC)
Amazingly this is all coming from a single IP address, according to Domas:
It is not "wrong" - someone is actively refreshing that page all the time. All requests are coming from single IP, and have referer header pointing to the product website. I guess they're inflating their stats that way :-) Domas Mituzas (talk) 08:47, 13 March 2012 (UTC)
Hi, I'm a host with WikiProject Teahouse, which is a supportive social and help space for new editors. The Teahouse is in pilot phase, and one of the main tasks of this period is reaching new editors to come and use the project. A helpful database report is available which creates a daily list of new but active editors (1 day, >10 edits; 4 days>20 edits). These lists are used for hosts to leave welcome and invitation messages on user talk pages.
The problem is that checking each page to see if there's been an invite previously made significantly slows down efficiency, and with 25 hosts aiming for a minimum of 500 invites per weak, that adds up in time costs. What would be amazing-great-phenomenal, is if someone with bot coding experience could whip up a simple machine to check the talk pages of the users on the database report, search for the word Teahouse, and then update an 'invited' column in the database report at regular intervals. I don't know how much time it would take, but it seems to be on the less complex end of things. Any help would be greatly appreciated by me, as well as the other Teahouse hosts. Or tips on who might be able to make such a thing. Thanks so much for your help and suggestions! Cheers, Ocaasit | c02:16, 13 March 2012 (UTC)
I was entertaining an idea that articles could have a custom search which allowed readers to dig through all of the references on the page, or all of the wikilinked articles. It just so happens that Google has an 'on the fly' custom search engine which can do just this (google.com/cse/tools/create_onthefly) but it's blacklisted. What are the issues involved here? Would a custom reference/wikilink search have to be open source? Could the Google custom search be taken off the blacklist? Is there interest in providing this type of focused search for articles? I personally think it would be a good asset for readers and editors as well, to help with research and verification. What do you think? Ocaasit | c17:52, 13 March 2012 (UTC)
I'm not planning to go around using it on articles without consensus, but it seems like the block was for an unrelated reason. How can I go about getting an exception to test this out (starting on my talk page). Ocaasit | c18:24, 13 March 2012 (UTC)
<strongclass="error"><spanclass="brokenref">Cite error: There are <code><ref></code> tags on this page, but the references will not show without a <code>{{Reflist}}</code> template or a <code><references /></code> tag; see the <spanclass="plainlinks"><aclass="external text"href="//en.wikipedia.org/wiki/Help:Cite_errors/Cite_error_refs_without_references">help page</a></span>.</span></strong>
I've just WP:PURGEd it (link) and it displays OK now. The error message is still there at the bottom, it's because one of the sections has some <ref>...</ref> but there is no {{reflist|close=1}} in the same section. --Redrose64 (talk) 20:59, 13 March 2012 (UTC)
Gadget Clock disappeared
The clock was visible earlier today. I just noticed it's totally disappeared from the top of my page. Is this gadget being worked on? I'm using Modern skin. Maile66 (talk) 21:07, 13 March 2012 (UTC)
Is it just me or is the Commons and all media on it really slow right now? And it looks like images not on Commons but on Wikipedia are slow as well? And when I Google "commons" the first result is this? Gary King(talk · scripts)02:44, 4 March 2012 (UTC)
Does it appear as the first result for you as well, or is it just me? In all my years of using Google, I don't think I've ever seen a result so irrelevant... (even though it's obviously related in being on the Commons). Gary King(talk · scripts)18:38, 12 March 2012 (UTC)
I <3 the numbered static column for use with sortable tables. I want to make a sortable table with a single thumbnail image of a map that takes up an entire static column (e.g., using rowspan) or even just make the image stick beside the table while still being accessible. Here is the example i've worked on: List of Perth suburbs#City_of_Armadale. I want it to include postcodes later on. I think if the numbered static column; which is complex; works then this should be easier. Warmest Regards, :)—thecurranSpeak your mindmy past08:49, 13 March 2012 (UTC)
This won't work. When you set the thumbnail's column to have class="unsortable", and you go to sort one of the other columns, the image ends up being duplicated. It would seem to me to be easier to include the image outside the table. --Izno (talk) 14:13, 15 March 2012 (UTC)
I forgot 70000 above. Just an example of how messy this solution is, and I don't know whether it's stable if EasyTimeline is altered. PrimeHunter (talk) 16:32, 15 March 2012 (UTC)
See my contributions from 01:48 until 01:51 on 16 March 2012. For all of these edits, I clicked the edit tab at the top of the "See also" section, and whenever the /* See also */ code was present, I was careful to preserve it, but it never appeared in the first place for many of these edits. Any idea why not? I'd noticed this happening earlier, and I expected that one of the developers was modifying something or that someone was editing a relevant MediaWiki page, but because I opened the pages in the order that I edited them (bigtime tabbed browsing), I doubt that's the case. My skin is Monobook; I'm running IE8, although I doubt that's relevant. Nyttend (talk) 01:57, 16 March 2012 (UTC)
For another example: I just edited the see-also sections for the thirty-one pages whose links are italicised in the bottom three rows of the navbox at the top of National Register of Historic Places listings in Indiana (Orange through Whitley), and of these thirty-one, just ten (largely scattered through the group) displayed the /* See also */ code. Nyttend (talk) 02:05, 16 March 2012 (UTC)
I've just had a look at your contribs list, picked a few at random, gone for the [edit] link at the top of "See also" - and without changing anything, I noticed that the edit summary is already blank. This suggests that the MediaWiki software is somehow unable to detect the heading text for the pre-fill - which is the opposite of a problem which I had some months ago (an edit summary which should have started off blank was being pre-filled with something that had been mis-detected). The problem then was that there was a hidden section heading within the section that I was editing, and the software was assuming that was the real heading; so somehow we're getting the opposite here - what looks like a real == See also == heading is not being detected as a heading. --Redrose64 (talk) 10:06, 16 March 2012 (UTC)
OK. I've just gone into one of these "See also" sections, blanked the line with the heading, re-typed it in afresh and saved. The page diff shows that there has been a change, so my guess is that there was some kind of nondisplaying Unicode character(s) on that line, which caused the MediaWiki software to fail to detect the heading. --Redrose64 (talk) 10:20, 16 March 2012 (UTC)
If searching could have an option to limit to Good and Featured Articles, that would seem consistent with an early design goal of Wikipedia. As I understand matters, an early plan was to copy the better articles to something called Nupedia.
My personal view is that GA/FA is not a guarantor of better research, because the reviews (what little I've seen of them) seem to focus on writing, format, and the like. It's also important to consider whether such a limit would reveal large gaps in coverage, rather than show better coverage, in most searches. And as the encyclopedia has more articles, now approaching four million, more searches will give results that satisfy visitors, and a limit might need to default to fuzzier searching to produce similar usefulness.
So I tentatively propose one solution: That search results show whether a result is a Good Article or a Featured Article. The label can have an icon but it would show this status in words. That way, the search itself would be done in the familiar easy way and the results would be compiled at your servers in the usual way.
I've just found that there's at least one other class of articles (A or A+ or some such) that might be included in this proposal.
This is not for the advanced-search page. This is for most visitors, and I think they mostly don't do advanced searches. But it might be useful that when advanced-searching in the article namespace an option to search only for better articles would be by a checkbox.
As an extra feature, each results page could have a button that would modify the search to add GA/FA as a criterion. That would be useful when results are very numerous and some way to limit the results would be helpful.
File:Pi-unrolled-720.gif, which is used in thumbnail form at the top-right of the article Pi as well as at Mathematical constant#Archimedes' constant π, is not showing up in either article. I've tried bypassing my browser's cache and purging both articles, as well as purging the image description page for the image both here on Wikipedia and at Commons, to no avail. The image displays fine at its description page. I am using Internet Explorer 8 and the Monobook skin. Any thoughts? Thanks. —{|Retro00064|☎talk|✍contribs|} 02:53, 16 March 2012 (UTC).
Thumbnails of animated gifs often won't have the animatd aspect work due to wiki parsing. As the first frame of this is blank (as far as I can remember) when parsing it will freeze as a blank image. PuppyOnTheRadiotalk04:25, 16 March 2012 (UTC)
But I am seeing a red X inside a box at the upper-left corner of the space for the image, which is the image that the browser displays in place of dead or non-working images. Could what you described cause this? Thanks. —{|Retro00064|☎talk|✍contribs|} 05:49, 16 March 2012 (UTC).
I'm not seeing either the animation, a thumbnail still from that, or a red cross "bad image" icon. What I see is a thick vertical bar in grey - which might possibly be the border around a zero-width image. Before this appears, there is a white rectangle of the expected size, but that collapses after a second or so. Windows XP, Firefox 3.6.28, Monobook. --Redrose64 (talk) 10:11, 16 March 2012 (UTC)
Seems to be working now. Commons lag or something, perhaps? Also, animations not continuing to work in resizes and thumbs appears to just be a Wikia bug; Wikimedia and other more stable projects do animate resizes and thumbs of animated images properly. — Isarra (talk)23:13, 17 March 2012 (UTC)
Well, any site that wants to live on the bleeding edge can follow SVN HEAD, and any site that wants the same version running Wikipedia can use the revision (from the wmf branch) specified on Special:Version. No one has to stick with the official releases. Anomie⚔01:51, 19 March 2012 (UTC)
length of edit-summary box
Hi, any reason the box can't be changed from the current ~80% to 100% of the width of the edit box? Would it be a trivial change of number in the software? Is it worth filing a bugzilla for this? Tony(talk)03:29, 17 March 2012 (UTC)
I have tried this: and found that it also affects the width of the "Subject/headline" window that you get when starting a new thread. Without the css above, the window sits to the right of the "Subject/headline" prompt, but with the css, its extra width forces it onto a new line. --Redrose64 (talk) 00:25, 18 March 2012 (UTC)
Yes, I've just seen this. In fact, the width of both subject/headline and edit-summary boxes exceeds that of the edit box by a few percent, it seems, when 100% is chosen. That's weird, although the more the better for the edit-summary box (we do now have the option in prefs to be allowed long edit-summaries—this suggests the need to up the width of the edit-summary box for users who choose this option, and to separate out the coding for the widths of the subject/headline and edit-summary boxes. How much hassle would that be? Is it worth my chiming in at the bugzilla that has stopped short of wider adoption? Tony(talk)01:23, 18 March 2012 (UTC)
IIRC, they basically adopted the gadget into core, didn't they? The reason for the whole mess is that the limit actually is bytes (due to the underlying MySQL database field limitation), while the standard browser mechanisms limit by characters (which in UTF-8 may be up to 4 bytes long). The old limitation of 200 bytes while the real limit is ~250 bytes tended to work well enough in most European languages that use mostly ASCII characters (1 byte each) with a few accented characters (generally 2 bytes each) thrown in, but routinely caused truncation for Chinese and other languages where most characters are multiple bytes. Anomie⚔01:59, 19 March 2012 (UTC)
It happened when I clicked [edit] on Ahalya. Also when I tried clicking on "templates for discussion" and "Template:Infobox Hindu deity" link on the top of the page. --RedtigerxyzTalk19:22, 17 March 2012 (UTC)
It's a known issue - certain types of vandalism attack will turn the entire page - even the blank areas - into a malicious link. Since this whole-page link is entirely transparent, and covers up all the text, any valid links are covered by it as well, invisibly. As a result, when you click on what looks like a valid link, you're actually clicking on the whole-page link. --Redrose64 (talk) 19:47, 17 March 2012 (UTC)
Hang on, that explanation doesn't make sense. The edit to Template:Notelist was reverted on 18 February; this report was submitted on 17 March. Could you tell us which other articles were taking you to the site? It Is Me Heret / c14:09, 18 March 2012 (UTC)
If he was logged out, then there is a known issue where pages that redirect to another page will sometimes redirect to an old version of the page. ---— Gadget850 (Ed)talk14:49, 18 March 2012 (UTC)
I've noticed in various articles there is a difference in the font-size of section and sub-section headings. Section headings are displayed in one size, sub-section headings are then displayed in only a slightly smaller size. For some strange reason, sub-section headings also appear to be bolded, (at least on my screen they are). This bolding gives the (false) impression that sub-section headings are the same size or even larger than section headings.
To take but one example, on the HMS Duke of Edinburgh (ship) page, which is a fairly small article, the 'Description' section and the 'Armament' sub-section makes them very difficult to tell apart when they are both visible at the same time. It's only on reading the various numbers that the situation becomes more clear.
I would have thought that it would be a lot better if section headings were bolded and substantially larger than sub-section headings. You would then have a far more logical, descending order of sizes and everything else.
My userpage contains many userboxes. While I use a fairly fast broadband connection which can load the page with ease, I attempt to be considerate of people in rural areas and those who can't afford broadband. I recently went through my userboxes (previously there were ~120 of them) and deleted those that were unnecessary. I still feel that I could be more considerate, however. Is there a way to make my userboxes section empty by default, requiring you to manually expand with a button in order for it to load?--YutsiTalk/Contributions16:20, 18 March 2012 (UTC)
It is possible to create expandable sections using the {{collapse top}} and {{collapse bottom}} templates. However, I don't think this is what you want, as the collapsed content will still load with the page; it is just hidden from view until clicked. The simplest solution is probably to place all your userboxes on a subpage (e.g. User:Yutsi/userboxes), then link to that subpage from your main user page. – PartTimeGnome(talk | contribs)16:39, 18 March 2012 (UTC)
This is driving me crazy. The image File:OSU Numbers Garden.JPG is used on two pages, including Peter Orno. If I view those pages when I'm not logged in, the image displays properly. If I'm logged in, I don't see it. This is not a problem with images in general; it's specific to this image. It's happening in two different browsers: Firefox and Google Chrome. I've closed the browsers, reopened them, restarted the computer, changed my preferences, etc., but it still happens. I have not changed my skin, however -- I'm using Monobook. What could possibly explain this peculiar display behavior? --Orlady (talk) 01:18, 19 March 2012 (UTC)
My thumbnail size setting is 250px, same as it has been for several years. Why should that affect the display of this one particular image? --Orlady (talk) 01:41, 19 March 2012 (UTC) I've experimented on that page. It appears that if I force the width parameter to be any size other than 250px, the image displays, but it does not display at 250px. I tried 150px, 200px, 220px, 249px, 250px, 251px, and 300px; it displayed at every width except 250px. --Orlady (talk) 01:48, 19 March 2012 (UTC)
The 250px thumbnail of that particular image is corrupted. I've tried purging it (go to the image description page and add ?action=purge to the URL, or for Commons images use the "purge" tab on Commons's image description page), which seems to have fixed it. Anomie⚔02:09, 19 March 2012 (UTC)
It's a recurring problem that images sometimes fail at certain sizes so I guessed you might have different sizes when logged in and out. I see it at 250px but it may have been after Anomie purged. PrimeHunter (talk) 02:15, 19 March 2012 (UTC)
Java Scripts do not work any more since MediaWiki upgrade
Since Mediawiki has been upgraded to 1.19, my user defined javascript does not more work. I had added some edit buttons that I found very useful. Since Mediawiki has been upgraded to 1.19 they may appear or not and now mostly not.
My java console doesn't retrieve any error.
Project : eo.wikipedia
Version : 1.19wmf1 (r113176)
OS : Windows XP
Browser: Firefox 11.0
I submitted this issue in bugzilla (35154), and was answered it is not the right place: "bugzilla isn't really for end user technical support like this.."
First thing I notice is the use of addOnloadHook();, which has been deprecated. Use $(document).ready(function() {...} ); instead and see if that resolves the problem. — Edokter (talk) — 22:34, 18 March 2012 (UTC)
You shouldn't need either an admin or a bot. Go to the project banner sandbox page, and you will see two boxes below the project banner. The second one has a bunch of redlinks, each one followed by a blue "create" link. Click that "create" link, this will open an edit window which should be pre-filled like this:
Fill in |project= so that it becomes |project=Cryptography; leave |topic= alone. You shouldn't need to make any other changes, and the edit summary is pre-filled, so click on "Show preview" to make sure, then save it. Do this for each of the cats. --Redrose64 (talk) 18:22, 19 March 2012 (UTC)
Oh! I was under the impression that for each of the quality assessment categories there were further subcategories for the importance ratings to create, but now I realize that was a misunderstanding. So instead of 55 categories I only need to create 11 by hand. That's doable. Thanks for the help! Nageh (talk) 19:09, 19 March 2012 (UTC)
Missing deleted edits
Laura Ramsey is/was tagged for deletion per G4. I went to check to see if it is similar to what was deleted after an AfD, but as Special:Undelete/Laura_Ramsey shows, there aren't any deleted edits showing up pre 2012, even though the AfD was in September 2011. Am I being stupid and missing something, or is this a bug? SmartSE (talk) 18:47, 19 March 2012 (UTC)
Could you please record any issues with Wikipedia since the rollout of 1.19 here, so that we can keep them all in one place, please. Thanks in advance, BarkingFish00:06, 1 March 2012 (UTC)
Any way to get the green back in diffs?
Resolved
I see the new diff system was implamented, possibly in the last few minutes. It's nice being able to see tiny changes better but I really like the old green color of the right column a lot better as it's much easier to read. Any way to get it back? I assume it'd have to be css but if so, then so be it. ♫ Melodia Chaconne ♫ (talk) 23:12, 29 February 2012 (UTC)
If you want to change only the background color of the right column, add the following code (adapted from MediaWiki:Gadget-ClassicDiff.css, which should be implemented as a gadget soon) to your monobook.css:
The diff color change has been reverted, and the reversion has been synchronized to the Wikimedia servers. The yellow/green coloring should be back now. PleaseStand (talk) 02:09, 1 March 2012 (UTC)
HistoryNumDiff problem...
Resolved
I assume the change in colors in diffs has broken this somehow. I have HistoryNumDiff enabled in my preferences and now it's showing the changes double, like so:
(cur | prev) 22:58, 29 February 2012 The Bushranger (talk | contribs | block) . . (+234) (+234) . . (→Unfounded sanction and possible admin tools abuse: arbchive) (undo)
The gadget HistoryNumDiff adds the net diff change to the history. MediaWiki 1.19 has just been rolled out to the English Wikipedia and has this feature built in (see mw:Special:Code/MediaWiki/111800). I would recommend disabling HistoryNumDiff in your gadgets. Goodvac (talk) 23:37, 29 February 2012 (UTC)
I've had an issue with this, I have always had HistoryNumDiff unchecked in my preferences, but suddendly today, coninciding with this apparent software change, now it has started appearing without having checked it. When I check the option, the HistoryNumDiff figures appear in duplicate, while unchecked it appears only once; there appears to be no way to get it to sling its hook and shove off. Any suggestions on how to get it to vanish, not to appear once, twice, or once again, but no appearences at all? Kyteto (talk) 01:40, 1 March 2012 (UTC)
MediaWiki 1.19 automatically includes the net diff change and the total bytes listed in the history. If you want to get rid of the net diff change, add the following to your skin.css:
I think the best thing would be to rewrite HistoryNumDiff to provide a "±" switch between size and dif size. — AlexSm01:53, 1 March 2012 (UTC)
Well I've pasted that line of code at the location provided, I'll wait and see if it makes a difference, no change so far even with a cleared cache. But doesn't it suggest that there's a deeper flaw, what could be the possible point of 'doubling up' on the figure? And shouldn't the GUI box to switch off feature A...actually switch off feature A, rather than turn on and off a redundant/pointless duplicate? If that how it's supposed to work/is it working like that for other people/am I simply stupid? Kyteto (talk) 03:09, 1 March 2012 (UTC)
Apparently there was a function named akeytt() in the site javascript, which has disappeared. I had to delete a call to it from one of my user scripts to make things work again. — Carl (CBM · talk) 01:22, 1 March 2012 (UTC)
User:Animum/easyblock.js is one commonly used script that calls the function, and I suspect that it is the cause of a report that one of my scripts was not working. It took me a while to identify it as a problematic script because it does nothing for non-admins like me. PleaseStand (talk) 02:02, 1 March 2012 (UTC)
Actually akeytt has been deprecated a little longer, since 2009. See also mw:RL/MGU#ta[] Tooltip and Accesskeys. Back then this tooltip/accesskey feature has been integrated into core MediaWiki for all wikis to use and to enjoy localization as well.
Since MediaWiki 1.16 (or earlier) we've replaced the deprecated variables (the global ta object and akeytt function) with dummies to avoid scripts from throwing exceptions for undefined variables. After 3 versions they were eventually removed in MediaWiki 1.19. Note that the errors appearing for scripts still using them are not breakages of functionality. The scripts currently referring to ta and akeytt haven't done anything since 2009.
The addPortletLink function takes a parameter for 'tooltip' and 'accesskey' for scripts that add new links and want to use tooltips and/or accesskeys. To modify the default tooltips and accesskeys for portlet links outputted by default (e.g. "What links here", "My preferences", "View history" etc.) change the interface messages in the MediaWiki-namespace (tooltips / accesskeys. Krinkle (talk) 03:27, 1 March 2012 (UTC)
I'm not 100% sure if this is due to MW1.19, but I think it might be. Anyway, when I'm looking at an old revision of a page there are edit section links available, however when I click on one it takes me to editing the same numbered section, but of the current revision, not the old revision that I was looking at (does that make sense?). So, is this a new thing or has it always been like that? Jenks24 (talk) 01:37, 1 March 2012 (UTC)
Roan thinks the other two are probably also issues with not using HTML5 yet, but you can file bugs if you'd like. --MZMcBride (talk) 03:17, 1 March 2012 (UTC)
This is indeed just a note in the validation report. The HTML5 standard does allow it, and afaik browsers never had a problem with it. These empty portlets are also (albeit being invisible anyway) additionally hidden through CSS so screenreaders shouldn't have an issue with them either (note that this is not new in MediaWiki 1.19). Krinkle (talk) 03:32, 1 March 2012 (UTC)
I seem to recall a very similar problem just over a year ago (1.17), but at the time there were three errors: the <html class=> and the two </ul>. I remember commenting that you couldn't use a class= attribute until some classes had been defined, and that classes can't be defined earlier that the <head>...</head> section, which is always enclosed by <html>...</html> and so any classes are invisible to the <html> tag.. --Redrose64 (talk) 12:33, 1 March 2012 (UTC)
It's valid in HTML5 though. It also does have a function, it's used by a javascript that (in theory) will actually run before the body might be in the dom model. Anyway that is something that have to be further tested, but the ResourceLoader is complicated enough, i'm not touching it ever :D —TheDJ (talk • contribs) 21:49, 1 March 2012 (UTC)
While the times in the list itself is fine, the part in the box (following "in the last 72 hours, as of") shows UTC time now. I have to hope this is a bug, not a feature. ♫ Melodia Chaconne ♫ (talk) 02:46, 1 March 2012 (UTC)
I have deployed the bug fix. Please note that on enwiki UTC is intentional, see the MediaWiki:Wlnote. Any sysop can replace it with the default MediaWiki message by just deleting that article. Unlikely I will come back here, so please follow up on T36835Hashar (talk) 12:51, 1 March 2012 (UTC)
Didn't it always show UTC time? But I've taken Hashar's suggestion and deleted the message to use the MediaWiki default that uses the users local time. Anomie⚔23:31, 1 March 2012 (UTC)
It most certainly did not use UTC time. I have to wonder why the hell it'd be "intentional" on enwiki. I've been using it ever since I started using my watchlist to see when I last checked it, based on my own time. I see it's be half fixed, as it now appears AFTER the date instead of before. ♫ Melodia Chaconne ♫ (talk) 03:23, 2 March 2012 (UTC)
Ah, I found it! In 1.18, the watchlist used MediaWiki:rcnote for time periods over 1 day, and that message behaves just as you describe. In r102284 the watchlist was changed to use MediaWiki:wlnote only; you may also notice that the watchlist is now displaying "in the last 72 hours" rather than "in the last 3 days". Anomie⚔04:21, 2 March 2012 (UTC)
Yeah it seems to be normal now, outside the 72 hours instead of 3 days (why the change? Not a huge deal though). On a side note I see the missing edit button when viewing diffs issue has finally been taken care of as well. ♫ Melodia Chaconne ♫ (talk) 05:03, 2 March 2012 (UTC)
"No longer working" isn't very descriptive. What exactly isn't working? Where are you importing this? What browser are you using? What behavior do you expect? --MZMcBride (talk) 03:06, 1 March 2012 (UTC)
Any way to remove characters added / removed from Contributions pages?
It's super unsightly and makes the article titles, the far more important item, form a ragged edge. Especially for lists of my own contributions, I do not care how much I added or removed; I wanted to be reminded of what pages I've been looking at lately. Is there a way to hide them just from Contributions pages? Or, at the very least, align them in their own columns, and return to aligning all the titles (perhaps with a "minor" marker) together? SnowFire (talk) 03:07, 1 March 2012 (UTC)
You can also add the code .mw-special-Contributions .mw-plusminus-pos, .mw-special-Contributions .mw-plusminus-neg, .mw-special-Contributions .mw-plusminus-null { display: none; } to your skin.css page to hide them, though the dots around the count and a space appear where it would be. - Purplewowies (talk) 03:35, 1 March 2012 (UTC)
Sorry, not that familiar with skin.css or the required format. Do I need to create a file called skin.css (in addition to monobook.js), or should it be monobook.css (which is what appeared when I clicked the link above)? Also, does the CSS code need to spread across several lines, or entered on one line?
Like SnowFire, I don't like the bytes changed disturbing the alignment of the article names. Could they be moved to the end of the line? Astronaut (talk) 11:47, 1 March 2012 (UTC)
Is anyone having issues with the older toolbar? When I go to the edit window, sometimes some of the buttons are missing. The only button that has appeared every time is the "cite" button (which, when I'm having this problem, is often the only button to come up). This just started happening tonight. I'm using IE8. - Purplewowies (talk) 03:41, 1 March 2012 (UTC)
This is not limited to IE. I'm using Firefox (versions 9 and 10, on different machines). I'm consistently seeing only about half of the buttons on the older toolbar. --Orlady (talk) 03:50, 3 March 2012 (UTC)
I've had it on various versions of Safari for more than a year. The best thing to do is to chose "show edit toolbar" in preferences and swap out the old toolbar. Truthkeeper (talk) 04:00, 3 March 2012 (UTC)
Firstly, I don't really like the new namespace selector in the move page function. I don't see the benefit of it and it'll confuse newer users who don't know what a namespace is. However a more serious issue is that the revert button in the move page feature doesn't work at all; it produces a "No such target page" error message. I just tried it at User:Graham87/test, but it also didn't work at User:Malcolm Farmer, where I was trying to use the move page feature to import old edits from the Nostalgia Wikipedia. Graham8705:59, 1 March 2012 (UTC)
The change was necessary to fix bug 29454 and the reasons are explained on rev:110209.
I don't think this error is related to the new namespace selector because the (old) url parameter for full title still works fine (example) Krinkle (talk) 00:53, 12 March 2012 (UTC).
I've added some additional test results and information to bugzilla:34887. New title: "$3 and $4 are not being substituted in {{urlencode:}} in message "movepage-moved"". Krinkle (talk) 01:49, 12 March 2012 (UTC)
Contributions lists for truncated IP addresses don't work anymore
The contributions lists for truncated IP addresses used in old versions of Wikipedia's software (e.g. 62.253.64.xxx, which made this edit that I just imported to Malcolm Farmer's user page), and 11.105, which is used as an example at User:0, do not come up with any contributions. This is also true in the former case at the Nostalgia Wikipedia. Interestingly, the deleted contribs lists for these addresses come up fine. Graham8705:59, 1 March 2012 (UTC)
Popups (which I assume work via AJAX API calls) provide the full list, which is suggestive of a problem with Special:Contributions itself rather than the underlying functionality. But that would be weird in a sense, because you'd have thought it was a problem handling "unusual" usernames in general. - Jarry1250[Deliberationneeded]10:32, 1 March 2012 (UTC)
Hey all. I was wondering if someone could replicate an edit conflict I just had, where I swear the intervening edit didn't display on the page at all, so when I copied my addition from the bottom box to the top, I obliterated it. - Jarry1250[Deliberationneeded]10:43, 1 March 2012 (UTC)
Can't replicate, but a very similar thing happened to me a few days ago. Ended up making a right old mess what with edit conflicts. Astronaut (talk) 11:40, 1 March 2012 (UTC)
When I get an edit conflict, I don't attempt to amend the upper window; instead I mark and copy my new post, back out and edit the section again, paste in my new text and save (as I advised to somebody else on 26 January 2012). --Redrose64 (talk) 14:01, 1 March 2012 (UTC)
Yes, well either way it's clearly defective to have diffs that claim to display one thing but actually display another (thanks for restoring the other edit I managed to lose, btw). - Jarry1250[Deliberationneeded]
Editing old revisions: Wrong "changes"
Resolved
When editing an old revision (such as this) and checking the changes to the current revision via "Show changes", what is shown is the changes to the old revision, not to the current revision as claimed. Huon (talk) 15:25, 1 March 2012 (UTC)
== Unexpected results using [Show changes] ==
Profile
Chrome 16.0.912.63 m
Monobook
Upon logging in for today's session, I noticed a number of stylistic changes while editing... most are welcome. When editing an article from an older page-version however, the [Show changes] button no longer shows the live-dif. Is this something that can be tweaked? -- WikHead (talk) 13:58, 1 March 2012 (UTC)
The changes to the Short Pages report are quite awesome IMHO. But I do have one, in comparison minor, complaint. The report has lost it's header. The header, among other things, gave a link to the report's talk page. Now, there's no way to get from the report to its talk page. An example of how this header used to look can be seen on some of the other special reports like Special:DoubleRedirects. - TexasAndroid (talk) 15:39, 1 March 2012 (UTC)
So, in theory, if I create a new "header" at the link you gave, it should display at the top of the report? That would work out nice enough, I would think. - TexasAndroid (talk) 15:51, 1 March 2012 (UTC)
And I have added a similar new header for the Long Pages report, which appears to have been updated similarly to the Short Pages report. - TexasAndroid (talk) 16:07, 1 March 2012 (UTC)
Close. At a minimum I remember that it also had the timestamp of the last generation of the cached version, which I do not see on that generic one. Unfortunately, beyond that and it linking to the talk page, I do not remember exactly what was on it. - TexasAndroid (talk) 01:58, 2 March 2012 (UTC)
Large template not transcluding
Not sure if this is a 1.19 issue or what, but if you go to WP:AFC/S, you'll notice there is a link to Template:AFC statistics. If you look in the wikicode, however, you'll see that it should be transcluded. The template is quite large, not sure if that has something to do with it. —SW—spout15:46, 2 March 2012 (UTC)
Likely. On the discussion for one of the PC RFCs that are being set up, there was talk that there is a hard limit to transclusion. I'll see if I can find the conversation. - TexasAndroid (talk) 18:51, 2 March 2012 (UTC)
Looks like message MediaWiki:Logentry-move-move (and its variants) is now used both for logs and for history/contribs so pagemoves summaries are recorded like "U moved page A to B: some reason" (example). I think the username is really unnecessary in the edit summary, this does not go well with WP:Changing username and (probably) leaves less space for user added comment. — AlexSm17:16, 2 March 2012 (UTC)
===Page move text summary===
I noticed that since the 1.19 deployment, the move summary is now longer. Prior to the update, move summaries used "moved <Oldname> to <Newname>: <reason for the move>". Since the update was deployed, the move summaries now use "<Username> moved page <Oldname> to <Newname>: <reason for the move>". With the old method, the username is displayed in the logs and in the diff, just not in the edit summary. With the new method, the username is displayed in the logs, the diff, and in the edit summary. By forcing the extra <Username> moved page, it is further limiting the move rationale. Is it possible to use the old edit summary? Alpha_Quadrant(talk)
I've experienced two incidents in which an edit conflict occurred, but the second user was not alerted to the conflict and wrote over the first change. This seems to happen when the one user is editing the whole page and the other is editing just one section. The first incident involved a series of edits to Memorandum of Understanding on the Conservation of Migratory Sharks between 15:38 and 17:00 on 1 March -- I was editing a section of the article and the other user was editing the whole article, but for some time we weren't aware that we were edit-conflicting. In the second instance, my edit was to the whole page and the subsequent user's edit was to one section of the page. Ironically, in that second instance, the other user and I also conflicted on a talk page discussion of the item, but on the talk page we were both editing the same section, and I (as the second user) was alerted to the conflict. --Orlady (talk) 03:45, 3 March 2012 (UTC)
I have updated an ancient Bugzilla report (interestingly, one originally submitted by me) to include this new issue. --Orlady (talk) 04:12, 4 March 2012 (UTC)
I'm confident that this edit did not add 51,802 character and likewise I'm confident that this edit did not add 1,646 characters, both of which are currently being reported by my contributions (bottom of list). Either there's something counter-intuitive about what these numbers mean or there's a bug. Dpmuk (talk) 05:52, 3 March 2012 (UTC)
Wow, the diff sizes for your contributions towards the bottom are totally borked. For this edit where you removed text, the diff size is +9,183! If you look at the diff from the history view, by itself, the size is still +9,183, but with a few other entries, the size is correct (-242). Someone should file a bug for this. Goodvac (talk) 06:05, 3 March 2012 (UTC)
Never reported a bug before but now I've confirmed it's not me being an idiot I'm happy to do so. Would I be right in thinking mediawiki as the product and history/diffs as the component is the best place for this? Dpmuk (talk) 06:15, 3 March 2012 (UTC)
I've never filed a bug either, so I wouldn't know the process. ;) After reading mw:Bugzilla#Reporting a bug, I think you're right about the product and component. Also noting here that these inaccurate diff changes are occurring on older edits. The same phenomenon is visible at my earliest contributions. For example, this was a +4,250. Goodvac (talk) 06:27, 3 March 2012 (UTC)
Not sure what dates produce this behavior, but my earliest edits circa August 2010 seem correct. These ones appear ~2007. Chris857 (talk) 04:21, 4 March 2012 (UTC)
I've added this to the bug about page diffs. Basically it's not looking back to find the previous revision, but simply assuming there wasn't one, and it is a page creation. - Jarry1250[Deliberationneeded]23:24, 4 March 2012 (UTC)
I sporadically get a pop-up that says (I'm paraphrasing) that a script may have (or has) stopped running, and it asks me if I want to stop the script. Only happens on Wikipedia, and just since the rollout. I use Windows XP and my browser is Firefox 3.6.27 Maile66 (talk) 22:55, 5 March 2012 (UTC)
I also use XP and FF 3.6.27. The message is from Firefox, and is not necessarily related to Wikipedia: in Tools → Options → Privacy, I have "Remember my browsing history for at least" set to 91 days. The history does get very long, so periodically I open it (Ctrl+H), select View → by site, and zap those sites which I'm not interested in going back to (right click, select "Delete"). If I had visited a lot of pages in that site, then while zapping, it might throw the error "Warning: Unresponsive script A script on this page may be busy, or it may have stopped responding. You can stop the script now, or you can continue to see if the script will complete." which I believe is the one that you describe. I always go for "Continue", and it seems to complete OK. --Redrose64 (talk) 00:07, 6 March 2012 (UTC)
That's the exact message, all right. Except I have my browsing history set to only 10 days. I'll try cleaning out my history.Maile66 (talk) 00:32, 6 March 2012 (UTC)
Special:PrefSwitch
I was randomly checking the special pages and noticed Special:PrefSwitch is throwing:
when I'm logged in. I've already guessed it is related to bugzilla:25029 and noted it there, but I thought I should mention it here. (In fact I have no idea if this is actually new in 1.19.)
Mark Hurd (talk) 15:42, 14 March 2012 (UTC)
MediaWiki logo missing
It seems like ages since I've last seen it at the bottom of a Wikipedia page. The Wikimedia logo is there, and an invisible rectangle of the same size linking to the MediaWiki website is right next to it, but the logo itself is nowhere to be seen. At first I thought it was a glitch, but it's been months since then—probably more than a year, in fact. It could date back to the switch to Vector, but my memory isn't of much help here. (I use version 10.0.2 of Firefox, though the specific version seems irrelevant given the nature of the problem.)
A little warning: although I appreciate any replies which this may receive, I may be off-line for a few days and therefore fail to acknowledge them in a timely fashion. Waltham, The Duke of12:29, 10 March 2012 (UTC)
It's showing for me, both in Monobook and Vector. I wonder if you have some sort of ad blocker on your computer that is blocking the image. Anomie⚔13:10, 10 March 2012 (UTC)
I've already tried looking at the image file on my own, and could not. (And I've also looked at a page in Monobook, and the logo was still invisible.) I have now cleared my entire cache, and also added an explicit exception rule for Adblock Plus (though I couldn't see why it would block it). Problem's still there... Waltham, The Duke of23:53, 10 March 2012 (UTC)
While the latter image appears properly, the former does not appear at all; when I try to select it, I only see a tiny dot of a shade (which shows the absence of an image rather than, say, a white box). I can't see the copy either, and I didn't realise you had transcluded it here until I read about it. I've tried Internet Explorer 7 and the image remains invisible. This is utterly perplexing, especially considering that I am apparently the only one facing this problem (although I suppose it makes sense, because the issue would have been fixed a while ago if more people experienced it). Waltham, The Duke of13:27, 11 March 2012 (UTC)
They are identical: their size is 43 bytes and their dimensions 1 × 1. Looking at it this way, the problem doesn't seem to be browser-related—although, on the other hand, I downloaded the images through my browser, so... I've just had a look at a Wikipedia page on a friend's laptop, and the logo displays properly. And in my laptop it doesn't, and not just on Wikipedia but on Wiktionary and Commons. I cannot imagine what software might be causing this. Waltham, The Duke of14:10, 13 March 2012 (UTC)
Ah, now we are getting somewhere: I can only see the second image, which means it's the name that causes the problem. But why? It can't be Adblock; I've just de-activated it for Wikipedia and cleared my cache, and the problem persists (and, in any case, the image is also invisible in IE and there is no Adblock Plus there). Waltham, The Duke of16:25, 14 March 2012 (UTC)
There are some passive blocking programs like SpywareBlaster and Spybot's "Immunize" function, that can permanently filter addresses across multiple browsers ("permanently" meaning it's not a running program or add-on than can be disabled). It's all I can come up with to explain the symptoms you're describing. Have you ever run anything like that? Equazcion(talk) 16:37, 14 Mar 2012 (UTC)
Not that I can recall... If this is what's happened, though, you make it sound as if it cannot be reversed. I can live with not being able to see an image, but who knows what else such a program might do. Waltham, The Duke of17:35, 14 March 2012 (UTC)
It can be reversed, it's just not as simple as disabling a program. I'm not well-versed on all the places where passive filters are stored, but for starters you can check the registry key HKEY_CURRENT_USER\Software\Microsoft\Windows\CurrentVersion\Internet Settings\ZoneMap\Domains, and Firefox -> Tools -> Options -> Load images automatically -> Exceptions. You could also check your Hosts file in \Windows\System32\drivers\etc. If you see extensive lists in any of these locations, it probably means a spyware-blocking program was run that downloaded a blacklist and applied it to your computer. You should be able to clear entries easily enough. Equazcion(talk) 18:02, 14 Mar 2012 (UTC)
I've looked into the three locations you have suggested, and the second and third were empty. The registry location did have a series of files with the names of various websites, but there was nothing Wikimedia-related there; with the exception of Google and eBay, they were all websites I don't use, and most names looked like the kind to be found in advertising pop-ups. I therefore didn't see fit to tamper with the list. If you have more ideas, they are welcome at my talk page (or the Reference desk thread, as suggested below, if I remember to start it); if not, it doesn't matter. I'd like to thank both you and PrimeHunter for your trouble, and I hope the issue will resolve itself one day. What a silly little mystery, though... Waltham, The Duke of18:29, 20 March 2012 (UTC)
MW 1.19 developers broke texvc options and the recent update reduced available <math> output to either PNG or the raw LaTeX code. There is no HTML nor MathML anymore. I did not investigated these details before this event happened because there was a rumour that MathJax is ongoing and will supersede the texvc (which looks pretty reasonable). Of course, the decision to degrade the texvc before the advent of MathJax was foolish. But it will unlikely be reversed and now there is a problem to obtain an intelligent (i.e. touchable and copyable) output of numerous instances of <math> found in English Wikipedia. mw:Extension:MathJax is something of an unknown term of deployment in Wikimedia and, hence, we have to survive with little hope of cooperation with Wikimedia admins. Is there some ideas how to implement MathJax as, say, a personal script or gadget which can process the raw LaTeX (when texvc is disabled)? Or there is already a discussion about this problem? Incnis Mrsi (talk) 17:57, 18 March 2012 (UTC)
I am interested to know more about "stability", in terms of "stable version", i.e., how to quantify "stability" of an article (edits/week, article size variation, etc.) Is there a page, article, essay, tool, dealing with this theme?. --Best regards, Keysanger (what?) 20:41, 18 March 2012 (UTC)
After suggestions on the New Page Triage discussion page, we've opened a Request for Comment on adding the NOINDEX tag to unpatrolled articles - basically ensuring they can't be syndicated by google. If you've got an opinion or any comments, head on over there and post your two cents :). Okeyes (WMF) (talk) 13:08, 20 March 2012 (UTC)
Right, I see. The first three entries in the references list each have one superscripted letter (the last one of three or four) for which the backlink does not work. These are:
single instance of Template:refn in the reflist, in the order it appears on the text = same problem persisted.
single instance of refn, moved it to top of reflist = problem disappeared, everything ok.
first instance of refn moved to top of reflist, added 2nd instance of refn in the order it appears on text = no ghost ref links in the first 3 refs, but 1 extra ref link appeared in the original refn, as in Notes item 54, link c:abc [this is the current page]
first instance of refn moved to top of reflist. 2nd instance of refn anywhere in reflist but with its nested ref above it = 1 extra ref link appeared in the original refn, as above.
first instance of refn moved to top of reflist. 2nd instance of refn anywhere in reflist but with its nested ref below it = 1 extra ref link appeared in the original refn as above, plus cite error re: the nested ref of 2nd refn instance.
I don't know if it's an error with my settings, but on Chrome I cannot edit templates. Instead they appear in a grey box that I can't edit. I do not have this problem on Firefox. JDDJS (talk) 02:08, 19 March 2012 (UTC)
If you can edit Template:Seed then please name a specific template you can edit in Firefox but not Chrome, say exactly what you do to edit it in Firefox, and at which point it fails in Chrome. PrimeHunter (talk) 20:00, 20 March 2012 (UTC)
When I want to edit a reference template in article to be 30em instead of 22em, the template is in a gray box. I found out that if I click on the gray box and then press delete while my cursor is before the box, I can see the normal code and edit as I would normally. When I am logged out or editing from Firefox I do not see the gray box and instead just the normal code. JDDJS (talk) 17:11, 21 March 2012 (UTC)
This is still an issue, and it includes NOINDEXed user pages as well. Has there been a change to NOINDEX or the mobile site recently that could cause this? —danhash (talk) 16:04, 19 March 2012 (UTC)
See bugzilla:35233: the mobile site became indexed on about March 8. If it really doesn't support NOINDEX keyword then someone needs to file another bug request. — AlexSm20:51, 22 March 2012 (UTC)
Firefox update causing svg > png problems.
In the update to firefox 11, they changed the way that image files are displayed so that when an image is opened in full size, it is centered in the screen and has a grey background. This is good for jpegs, but many of our .svg files have a see-through background, meaning that images like File:Beta-aminobutyric_acid.svg appear like this when they are rendered by mediawiki as pngs. Is there anything that we can be do to solve this? If not us, then is there someway to change firefox's settings? SmartSE (talk) 20:33, 21 March 2012 (UTC)
(edit conflict) It's like any other browser except really dark. Most of them do a light greyish, but Firefox is... dark (this colour). Honestly sounds more like a Firefox is dumb problem than something that we could do much about here, but for now you could do what SD5 suggests. — Isarra (talk)23:41, 21 March 2012 (UTC)
Two solutions found from googling:
Technical: Expand omni.ja, found in Firefox's root program folder, using an archive program (7zip etc). Edit TopLevelImageDocument.css, search for "background-color:", and replace "#222" with "#FFFFFF".
Fixed There were two sets of stray double braces that interacted with {{cot}}/{{cob}}. When you would edit section 9, the following sectioos were in the edit window as if they were subsections. ---— Gadget850 (Ed) talk 19:30, 11 February 2012 (UTC)
Fixed Thanks so much, for the purge and the info! MJ (t • c) 15:30, 15 February 2012 (UTC)
Resolved
, 
Done, see here. --Redrose64 (talk) 13:46, 7 March 2012 (UTC)
Works for me too. Anyway, MediaWiki software changes (bugfixes or no) tend to go out to the Wikimedia Foundation sites first, of which English Wikipedia is one. Other users of MediaWiki often need to wait. --Redrose64 (talk) 23:34, 17 March 2012 (UTC)
You are invited to join the discussion at Help talk:Citation Style 1#RfC: Use "Vol.", "pp.", etc. consistently between citation templates, instead of ambiguous formatting like "9 (4): 7". The talk page at Help talk:Citation style 1 is where the discussion about most of our citation templates is centralized. — SMcCandlish Talk⇒ ɖ∘¿¤þ Contrib. 20:11, 19 March 2012 (UTC) — SMcCandlish Talk⇒ ɖ∘¿¤þ Contrib. 20:11, 19 March 2012 (UTC)
