This page contains discussions that have been archived from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.
I'm actually serious...the user links above the article view actually swaps from one side of my window to another when I try to mouse over it...This is happening on IE6 (which I'm unfortunately stuck with due to technical limitations on this computer, can't use IE7 or firefox). So...anyone know why that might happen?
Someguy1221 (
talk)
05:21, 8 August 2008 (UTC)
Of course not :-p It shouldn't need one, anyway...Regardless, well, my question is answered. Thanks brion; maybe when you're finally done making MediaWiki perfect you can fix Windows too ;-)
Someguy1221 (
talk)
01:44, 9 August 2008 (UTC)
The (semi) standard way is to edit each userbox so it has category suppression: see
User:Willscrlt/UBX/categories. Then edit the list to apply category suppression— you can also use {{ubxrow}} which does it automatically. I can help edit a userbox or two so you can get the idea. Category suppression should be applied to any template that adds a category. --——
Gadget850 (Ed)talk - 13:55, 8 August 2008 (UTC)
While I agree with that principle (although it would be nice if we could also all agree on what method of category suppression to use!), that doesn't solve the problem of ridiculous displays like the image you've got there :D.
Happy‑
melon14:04, 8 August 2008 (UTC)
My only qualm is that all of these are in userspace. BTW— there seem to be a number of userboxes listed at
Wikipedia:WikiProject Atheism that are not on the category list. It would be better to put all of them on a subpage and transclude them. Actually, it would be really nice if you didn't have multiple userboxes for the same concept, but I'm not going to belabor the point. --——
Gadget850 (Ed)talk - 14:13, 8 August 2008 (UTC)
I might get to the others later, but this should be easy. This is a different method- it adds the category only when the userbox is in userspace. --——
Gadget850 (Ed)talk - 19:26, 8 August 2008 (UTC)
I went through all of the templates listed there and added <includeonly>{{#ifeq:{{SUBJECTSPACEE}}|User|[[Category:Atheist Wikipedians|{{PAGENAME}}]]}}</includeonly> — is that what fixed it? −
Twas Now (
talk •
contribs •
e-mail )
08:55, 9 August 2008 (UTC)
Which, of course, works great, except when I enlarge the image using, e.g., the "ImageZoom" firefox extension. Then I can see only the left-most 150px of the image.
By moving the "width" spec from the "thumbinner" div to the "thumbcaption" div the page renders the same AND I can zoom the image. This seems a simple change.
I still have not ventured into the bot realm so I don't know about writing something that would accomplish this, so bear with me, please. I am hoping to port a travel orb template with images from another wiki; it consists of about 200 png images of country flags that have been photoshopped behind an orb face, sort of like a small round button. Uploading these files will take a while by hand, and I wondered if there was a method by which they could be uploaded en masse with the summary text being standardized for them? The filenames are the country names (not abbreviations), so the process could use the filename (minus .png extension) in the summary. Any ideas will be appreciated. Thanks!
Tigey Talk E-Mail 19:02, 7 August 2008 (UTC)
I would like that MediaWiki marks all talk pages with say "ns-talk" in the body tag of the rendered HTML page. Currently MediaWiki puts the namespace number as a class name in the body tag of the rendered HTML page. However there are many talk namespaces such as "Talk:", "User talk:", "Image talk:" and so on. Having the same tag on all talk pages would make it much easier to do namespace detection in CSS.
Longer explanation:
Many things need to look different when on different kinds of pages, that is in different
namespaces. For instance the background of the page is light blue on all non article pages, and message boxes are brown when on talk pages. There are several ways to detect namespace and adjust the looks of a template or a page. One way is to use
magic words such as {{NAMESPACE}} together with
parser functions such as {{#ifeq:}}.
Another way is to do detect the namespace by using CSS only, since MediaWiki adds the number of the namespace in the body tag in the rendered HTML page. For instance for this page the body tag looks like this:
The "ns-4" class means that this page is a page in the "Wikipedia:" namespace. This can be used in CSS code to detect namespace and change looks of something or even hide it based on what page the object is shown. For instance like this:
.ns-0.davidbox{/* Blue border when on "Wikipedia:" pages. */border:1pxsolid#0000FF;}
However, to detect if the page is some kind of talk page is messy, since there are many different kinds of talk pages. Then you have to do like this:
.ns-1.davidbox,.ns-3.davidbox,.ns-5.davidbox,.ns-7.davidbox,.ns-9.davidbox,.ns-11.davidbox,.ns-13.davidbox,.ns-15.davidbox,.ns-101.davidbox{/* Red border when on any kind of talk page. */border:1pxsolid#FF0000;}
If MediaWiki also marked all talk pages with say "ns-talk" in the body tag of the rendered HTML page, then it would be much simpler to detect a talk page in CSS. Like this:
.ns-talk.davidbox{/* Red border when on any kind of talk page. */border:1pxsolid#FF0000;}
That's all. And no, I don't have a MediaWiki Bugzilla account and I don't know how Bugzilla works, and I don't want to learn all that just to discuss one idea.
Thanks. And yeah, probably should be easy to implement. And would make namespace detection code in the CSS files much shorter and more readable. Having to add 9 class names for each actual talk page class when doing name space detection is one of the reasons I have avoided using CSS based namespace detection.
Thanks a lot SQL and Happy‑melon! I took a look at the discussion at
T17079, the code at for instance
[1] and the "Log Message:" at
[2]. Seems Happy‑melon suggested a nice little extension to my idea above and that SQL extended it even more. It seems SQL added this to the MediaWiki code: All talk pages will get the class "ns-talk" in the body tag. All non-talk pages, such as articles, "Wikipedia:", "Image:" and so on will get the class "ns-subject" in the body tag. And all "Special:" pages (see
Special:SpecialPages) will be marked with "ns-special" in the body tag. And as AlexSm noted above, the "ns--1" class and the "ns-special" class have the same meaning. But using "ns-special" in CSS code will be clearer, since "ns-1" and "ns--1" looks too similar but have very different meaning.
I agree with the extensions you guys did to my idea above. And as far as I understand this will be on-line the next time Wikipedia is updated to run on the latest MediaWiki code, right? (Such updates are done pretty often.)
Now I need to think about and experiment how to best use this new kind of namespace detection while still retaining the "demospace" functionality and the ability to lock for instance a message box to have the same looks in all namespaces. (I know several ways to achieve this, just need to figure out which way is the best one.) Anyone interested in this might want to watchlist
Template talk:Main talk other,
Template talk:Tmbox and perhaps
Template talk:Mbox since that is probably the pages where I will report my conclusions some weeks from now, and where we will discuss this.
Wow, that was quick! The "ns-talk", "ns-subject" and "ns-special" tagging is now live. I checked several different namespaces and all three of "ns-talk", "ns-subject" and "ns-special" work as they should. And I checked around, it seems to also be live in the other language Wikipedias and other Wikimedia projects like Meta and Commons.
I'm actually serious...the user links above the article view actually swaps from one side of my window to another when I try to mouse over it...This is happening on IE6 (which I'm unfortunately stuck with due to technical limitations on this computer, can't use IE7 or firefox). So...anyone know why that might happen?
Someguy1221 (
talk)
05:21, 8 August 2008 (UTC)
Of course not :-p It shouldn't need one, anyway...Regardless, well, my question is answered. Thanks brion; maybe when you're finally done making MediaWiki perfect you can fix Windows too ;-)
Someguy1221 (
talk)
01:44, 9 August 2008 (UTC)
The (semi) standard way is to edit each userbox so it has category suppression: see
User:Willscrlt/UBX/categories. Then edit the list to apply category suppression— you can also use {{ubxrow}} which does it automatically. I can help edit a userbox or two so you can get the idea. Category suppression should be applied to any template that adds a category. --——
Gadget850 (Ed)talk - 13:55, 8 August 2008 (UTC)
While I agree with that principle (although it would be nice if we could also all agree on what method of category suppression to use!), that doesn't solve the problem of ridiculous displays like the image you've got there :D.
Happy‑
melon14:04, 8 August 2008 (UTC)
My only qualm is that all of these are in userspace. BTW— there seem to be a number of userboxes listed at
Wikipedia:WikiProject Atheism that are not on the category list. It would be better to put all of them on a subpage and transclude them. Actually, it would be really nice if you didn't have multiple userboxes for the same concept, but I'm not going to belabor the point. --——
Gadget850 (Ed)talk - 14:13, 8 August 2008 (UTC)
I might get to the others later, but this should be easy. This is a different method- it adds the category only when the userbox is in userspace. --——
Gadget850 (Ed)talk - 19:26, 8 August 2008 (UTC)
I went through all of the templates listed there and added <includeonly>{{#ifeq:{{SUBJECTSPACEE}}|User|[[Category:Atheist Wikipedians|{{PAGENAME}}]]}}</includeonly> — is that what fixed it? −
Twas Now (
talk •
contribs •
e-mail )
08:55, 9 August 2008 (UTC)
why the interwiki bots are not patroling wikipedia:, template: namespaces?
They only patrol the article pages. For example check
Template:Split. The french version has only 4 interwiki links, and the english has 11 links. I can try to make a complete list with the pages from the "Wikipedia:" namespace. And then a bot can patrol the pages in that list and fix the interwiki links. But where should I ask for help? I will do the list, but I need someone to put a bot to work to patrol that list —
Ark25 (
talk)
11:59, 9 August 2008 (UTC)
The interwiki bots have a hard enough time building the web in the mainspace. Interwiki-ing the 'backstage' namespaces is definitely a lower priority. As-and-when we get a complete interwiki map of the article namespace, we can start worrying about the other namespaces.
Happy‑
melon14:21, 9 August 2008 (UTC)
That means something like "not in my life time". I don't expect the growing of the number of articles to slow down very soon, and the point where the other languages will have the same number of articles as the english wikipedia seems veeeery far away. Camon, at least the "Wikipedia:" and "Template:" namespaces should be considered to have more priority than decades or centuries away. In my opinion it is very important for a newcomer that starts editing to be able to inspect the other languages for templates, policies and guidelines. I can't believe that this issue can have such a low priority.
Ark25 (
talk)
14:48, 9 August 2008 (UTC)
Pywikipedia has or in any case had problems with interwiki links and noinclude tags in templates. Sometimes the interwiki links would be included in other pages. Therefore, you can't autonomously run an interwiki bot on templates. --
Erwin85 (
talk)
15:10, 9 August 2008 (UTC)
Ok I agree that on "Template:" namespace its a bit harder, but at least on "Wikipedia:" namespace should not be any problem. Also for templates should be a solution. Anyways I would be very happy to start with something. But I don't know where to find the right person to run a bot for a list of pages from the "Wikipedia:" namespace. I will take care of making that list.
Ark25 (
talk)
15:26, 9 August 2008 (UTC)
Not really easier as a rule as pages like
WP:AN/I have the interwiki links on a separate (and extremely bulky) header page. One reason for this, surely, is to put important stuff like interwiki links and noticeboard instructions somewhere where the archival bot can't find it and sweep it away as a stale section. —
CharlotteWebb16:33, 9 August 2008 (UTC)
Ok then, we will start with an "easy" list, with names of the pages like
Wikipedia:Deletion policy. One name per line. In case the page grows and the interwiki links will be removed and placed in a nasty format and so, the bot will notice that the interwiki links are gone, will skip the page, and will add the name of the page to another list - with the "buggy" pages. This second list will be inspected by someone, and that person will add the name of the format near the name of the page. Also that person will remove the name of the page from the first list. This second list will have to be patroled by a second bot (or the same, but in "smarter mode"), that will be wise enough to take care of the interwiki links that are somewhere else and not in the page.
Ark25 (
talk)
17:17, 9 August 2008 (UTC)
Waaaay down on the devs' priority list is a more efficient way of building interwiki links. I think in most ideas the linking is centralised: instead of building a
spider-web between every possible pair of articles, the interwiki links are recorded in a central table (probably on meta) and are dispersed in a 'star' pattern. This dramatically reduces the number of links that have to be made (in theory, from 23,000 per topic down to just 214) and enables the construction of the interwiki table to be centralised. So as-and-when that factor-of-100 reduction in workload occurs, we'll find interwiki linking much easier going. Who knows, we might then have time for other namespaces.
Happy‑
melon20:59, 10 August 2008 (UTC)
I guess you said 214 because there are 214 languages. And 23,000 because 214*213/2=22,721 (no. of interlinks in a star with 214 corners). I think:
The problem of interwiki links can be solved without developer's intervention. It can be solved by bots.
The number can be reduced from 214 to 1. How: For every page, we need an extra field that indicates where are the interwiki links located.
If that field is empty, that means the page contains it's own. interwiki links. The bot will look for them at the end of the page and adds there new interwiki links and gets them from there. If the bot doesn't find them, it will report the page so then a person can fix the problem (fills the value of that field mentioned above)
If the field is not empty, then its value points to the subpage where are the interwiki links. The bot is looking for the interwiki links there, adds new ones and gets the list of interwikis from there. If the bot doesn't find them, it will report the page so then a person can fix the problem (fills the value of that field with the correct name of the page where the interwiki links are, or voids the value of that field)
In the best case, that field will be part of the page (so then developers might be needed for implementing that). But, that field can be also an external page like "
http://interwikis.mediawiki.org/en/article_name-interwiki_page. So I think it doesn't have to wait for someone's priority list. I just need someone to volunteer to help by putting a bot to work. We can start with a very small list of pages, for example with 60 policies from there
Category:Wikipedia official policy, and test it untill we make sure it flawlesly adds the interwiki links at least to that small part of the "Wikipedia:" namespace. Then we grow the list, test it, adress the exceptions and problems and so on. You mentioned complete graph, star pattern, the number of the links and such. It was only your thinking there or there have been talks about it? I would like to see those talks if possible, if they existed —
Ark25 (
talk)
04:50, 11 August 2008 (UTC)
Which, of course, works great, except when I enlarge the image using, e.g., the "ImageZoom" firefox extension. Then I can see only the left-most 150px of the image.
By moving the "width" spec from the "thumbinner" div to the "thumbcaption" div the page renders the same AND I can zoom the image. This seems a simple change.
For some reason, instead of doing the redirect the edit summary says it should have done,
User:Redirect fixer replaced
XM and Sirius merger with the entire contents of
WP:FPChere. Odd thing is, in the page
history, the software says that revision was only 29 bytes (what it would have been if the edit went as expected). What the heck happened?
Hersfold(
t/
a/
c)19:29, 9 August 2008 (UTC)
This was nothing to do with the RedirectFixer script: one of the slave servers got its knickers in a twist and started entering corrupt edits. A number of edits, both bot and human, were screwed up before the devs took the server out of the network and blew it up or something (:D). IIRC Tim said they "might" be able to recover the original edits... meaning that it's down there just underneath "finish MediaWiki" on their todo list...
Happy‑
melon20:49, 10 August 2008 (UTC)
This is a technical question, although not really about Wikipedia's technical aspects.
How does the redirect page
■ work? I mean, I know how to type the symbol, but I didn't even know that was a valid web address.
Deamon138 (
talk)
00:19, 8 August 2008 (UTC)
Ah right, thanks! I knew there were ones like "%20" etc that do spaces and full stops or whatever, but I didn't know it could go that long. However, what happens if I wanted to go to "[[%E2%96%A0]]" instead of
■? Are percent symbols not allowed in article names or something?
Deamon138 (
talk)
17:49, 8 August 2008 (UTC)
Percent followed by two hex digits is not allowed, as it would be ambiguous. Same with an ampersand followed by some letters/digits and a semicolon. --
brion (
talk)
23:33, 8 August 2008 (UTC)
Sometimes it works, sometimes it doesn't, sometimes strange things happen: [[%20]],
0,
00, [[%E2]], [[%E20]], [[%E200]], [[%25E2]]. --
Carnildo (
talk)
05:13, 10 August 2008 (UTC)
And therefore [[%20]] fails for the same reason [[ ]] does (whitespace-only is invalid), and [[%25E2]] fails because it decodes to "%E2" (which is an invalid page name, as already discussed). —
Simetrical (
talk •
contribs)
21:22, 10 August 2008 (UTC)
So therefore it is impossible to make it go anywhere but square, as the hex names are impossible (on Wikipedia or on the web completely?). Okay thanks for the input everyone, this topic is another thing I shall be curious about now lol!
Deamon138 (
talk)
01:11, 11 August 2008 (UTC)
Any
Unicode character is valid in the path part of a domain name (the part after the domain name, like after "
http://en.wikipedia.org/"), although IIRC with a very small number of exceptions. Any "special" characters (including all non-
ASCII characters) usually need to be percent-encoded for actual use, like in an HTML href or src attribute, or when sending an HTTP request. But good browsers will tend to automatically percent-decode and percent-encode URLs to make them more readable, so that part isn't necessarily user-visible. In Firefox 2 you would have seen "%E2%96%A0" in the URL bar instead of the "■" you get in Firefox 3, for instance.
The domain-name part (before the third slash, after the protocol part like "http://") can't use percent-encoding. It uses a different system: see
Internationalized Resource Identifier for details. As IRIs are more widely supported, percent-encoding of international characters will no longer be necessary in all contexts.
URL may also be of interest. —
Simetrical (
talk •
contribs)
14:41, 11 August 2008 (UTC)
I posted on
User talk:HJensen 's talk page. I added more, and in doing so, I accidentally reverted a post he made in the intervening time. I have tried everything to restore his post and my explanation about the reversion, but nothing I post shows up on his page after I save it. My posts and his posts show up in the edit mode and in the history of his talk page, but not on his page after I save it. He is having the same problem!!!????? —
Mattisse (
Talk)
20:58, 11 August 2008 (UTC)
Editable edit summary when using the "new section" button
When you click on the "new section" button on a talk page, it automatically creates the edit summary with "SECTION_NAME: new section", which is okay. The problem is that there is no way to change this edit summary if you want to write in more details. Is it possible to have the "new section" button automatically generate the "SECTION_NAME: new section" edit summary, but then still display an editable edit summary to the user so they can write in a custom edit summary? —
OranL (
talk)
20:30, 9 August 2008 (UTC)
If you preview your edit before saving, an edit summary box is on the preview page, and you can add an edit summary before saving the edit. –
ukexpat (
talk)
18:19, 11 August 2008 (UTC)
Mine does not appear to be working like that. It just shows me the edit summary preview, and that always says "→SECTION_NAME: new section". There is not edit summary box anywhere on the page. —
OranL (
talk)
18:54, 11 August 2008 (UTC)
Oops my mistake - I saw the "Subject/headline:" box and thought it was an edit summary. I guess the only way to do what you want is to create a section by editing the full page. –
ukexpat (
talk)
19:01, 11 August 2008 (UTC)
In very long pages (like the
Policy section of the Pump), the one-pixel border around the page vanishes; the only exception is a piece of the line invariably surviving in the middle of the page, but in black instead of grey. I have Firefox 3, but the problem is not a new one and was also present in Firefox 2—as a matter of fact, this situation has existed for longer than I can remember. I've just thought of trying Internet Explorer, and it shows the border just fine. What is going on?
Waltham,
The Duke of22:16, 9 August 2008 (UTC)
I don't see this at
WP:VPP at the moment, but I see it at
WP:ANI. I'm using Firefox 3 on Linux, specifically Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.9.0.1) Gecko/2008072820 Firefox/3.0.1. I'll go ahead and report it to them. —
Simetrical (
talk •
contribs)
20:45, 12 August 2008 (UTC)
If I am wishing to make a template for my userpage to show various photos selected randomly, what syntax from this should I employ? Asenine 17:45, 11 August 2008 (UTC)
Thanks for that. Asenine 09:05, 12 August 2008 (UTC)
TOC and Infoboxes
While reading a discussion about infoboxes messing up well-designed lead sections, I was struck by the idea that since infoboxes like being on the right side of the page and TOC's seem to prefer the left side, there could be a happy place for both of them side-by-side. If the TOC was longer, the first main-body section would start below it; if the infobox was longer, the main-body would start under the TOC and wrap to the left of the infobox until it got past it, then go to full-width. (Or there would be another right-justified image, whatever)
The only way I could figure out for this was to do a table with __TOC__ in the first (left) cell and the {{infobox...}} in the second (right-side) cell. This worked out, except the table bit means that when the infobox is longer, the body text is pushed down below the extent of the infobox, i.e. there is a whitespace gap underneath the left-side TOC and the main text starts at full-width below the entire table extent.
Now I have no idea whether I've described the geometry properly, but if I have, does anyone have the code to accommodate this? TOC and infobox side-by-side, top-aligned under the article lead, article body text begins immediately beneath the last extent of the left-side TOC? Context is
here and associated talk. (Note: there be dragons - I'm asking more of a general question than at the specific article :) Thanks!
Franamax (
talk)
04:18, 12 August 2008 (UTC)
TOC top-aligned to the right of the article lead and to the left of any infobox would be nice too -- that'd be my preference, I think, so I could get to the TOC with the Home key. The seat of my pants says that there should be some way of doing this by diddling your User:<username>/monobook.css, but I'm a lousy css-diddler. --
Boracay Bill (
talk)
05:28, 12 August 2008 (UTC)
I'm actually looking for a solution for the anonymous reader here, i.e. coding for the default presentation of the article. I'm perfectly willing to torture my (or your) monobook.css until it gives the right answer, but I'd rather wait to see if the wizards can incant a more general solution...
Franamax (
talk)
06:06, 12 August 2008 (UTC)
Article body text. Article body text. Article body text.
As you can see it works fine. The trick is to place the boxes on the page in the right order and then use {{clearleft}}. Well, you are supposed to need the clearleft, but oddly enough it also worked without the clearleft in my example here. Anyway, here is a code example:
Article lead section text.
{{infobox}}
__TOC__
[[Image:Something.jpg|thumb|60px|A test image]]
{{clearleft}}
Article body text.
That's all there is to it. Oh, and don't bother about that the dotted <pre> box underlaps the other boxes, that is a known bug and a whole other issue. (Well, you can solve that by putting {{clear}} just before the <pre> tag thus pushing that box below any other boxes.)
Actually no, I want "Article body text..." to start immediately under the TOC and wrap between the left border and the infobox until it clears the bottom of the infobox, then go to full width. If I interpret your example correctly, the infobox extent is still determining the start of the main body text. I want the main body to be determined by the left-side extent, i.e. the TOC in all cases. (And of course it would actually be a bold section header with an (edit) button tucked in underneath the TOC).
Franamax (
talk)
02:07, 13 August 2008 (UTC)
I now took a look at this page in all three of my browsers. Firefox 2 and Opera 9 shows it like you want it. But my old Internet Explorer 5.5 shows it in the bad way that you seem to say that you see it now. So I guess you are using Internet Explorer and that even the newer versions of Internet Explorer shows it in the wrong way.
I did some testing. If I remove the image or move the image up between the infobox and the TOC then it works like you want it, even in my old Internet Explorer. Look at this:
Now it looks like you want it even in your Internet Explorer, right? Seems we have to first put all right aligned boxes, and then the TOC, and then the message body.
Yes that looks fine now, thanks! I forgot to mention I use Microcrap Internet Exploder. Now I'll have to try that technique but using the __TOC__ keyword rather than the class=toc you've used in your example.
Franamax (
talk)
05:21, 13 August 2008 (UTC)
I can't remember if I asked this before, but here goes. I'd like to know if there are any guidelines on minimum (or maximum, if applicable) screen resolution that should be catered for by Wikipedia pages. I normally edit with a usable browser area of approximately 1280x840 pixels, although my monitor can go to a higher resolution still. However I appreciate that there may be users with a lower screen resolution, and I believe this may cause images to float over the text in some cases. I'm considering adding some more images to the right of the list in
West Bromwich Albion F.C. seasons - at a width of 1280 there is easily enough room to do this, but at a width of 1024 it doesn't look feasible. I'd appreciate any guidance on this. Thanks. --
Jameboy (
talk)
12:11, 2 August 2008 (UTC)
Around half of all readers have a screen width less than 1280, and some of them would probably prefer not to have to maximise their windows. I would suggest there is not enough room at the side. --
zzuuzz(talk)12:17, 2 August 2008 (UTC)
Just for additional information, according to
W3 Schools, at
http://www.w3schools.com/browsers/browsers_display.asp , up to January 2008, ~86% users have screens with resolution of or higher than 1024×768, in order words, ~14% users have screens with resolution lower than 1024×768. That is about one out of seven users have screens with low resolution, and the actual figure may go higher as explained on the same web page, the web site is for people with interest in web technologies and so the average users may use screens with lower resolution. Furthermore, quite many people do not maximize their windows due to their preferences or other application running at the same time. Therefore, in my own opinion, pages should be designed such that it can display "properly" on a screen with resolution as low as 640×480. —Preceding
unsigned comment added by
Quest for Truth (
talk •
contribs)
19:21, 4 August 2008 (UTC)
In addition, according to
List of countries by number of Internet users, over 1/4 of US users don't even have internet access, with even higher numbers in the UK etc! What about them?
But, seriously folks, monitors don't last forever and when that CGA finally dies it will be replaced with a brand new <insert what will be last year's top of the line/this year's loss leader>. And then
2,489,407 pages are gonna look awfully chintzy.
Saintrain (
talk)
20:01, 4 August 2008 (UTC)
Hey, I don't mean that those people having screens with low resolution must be using old-fashioned machine. Actually there are quite a number of users of portable devices whose monitors can't have a higher resolution simply because the size of the screen is designed for mobility.
And I must repeat that many people don't maximize the window when browsing Wikipedia. Though there is no statistics and obviously it is difficult to have one, I'm pretty sure that it's not a small number as it is normal for people to have more than one windows opened in order to do multiple tasks or a complicated task involving more than one program at the same time. If one have a screen of 1280×960 and want to limit the space occupied by the browser to be a quarter of the whole screen (for example a notepad, a media player and an e-mail client are opened with the browser and the user want to keep an eye on all four window at the same time), the browser would have a humble resolution of just 640×480. (Actually why I suggest 640×480 in the previous comment is based on such calculation and I believe all assumptions behind are reasonable.)
Also, it is more difficult and inconvenient to view an "improperly designed" articles on a low-resolution screen, and you can try it by resizing your browser. On the contrary, it is easier for high-resolution screen users to adapt by resizing their browser. Advances are made quickly, and may be in the near future it becomes popular for us to have screens of millions or even billions of resolution, but right now it is better to be considerate. --
Quest for Truth (
talk)
18:39, 8 August 2008 (UTC)
I'm all for considerate and stuff, but something occurs to me, I don't ever remember seeing a complaint that articles were too big. Could you point to one? I've only seen concerns that other users may have a problem. Is this a real problem?
Saintrain (
talk)
00:23, 13 August 2008 (UTC)
Sorry. Looked at the article and talk and found no complaints about unreadability by the reader. Could you be more specific? Didn't notice the edit history summaries. But a list of area codes is a pretty unusual "article". Is that likely to be perused or just consulted? Can you point to a real article with complaints by the reader? A help desk request by the reader to make everything fit on an ipod?
Saintrain (
talk)
16:54, 13 August 2008 (UTC)
That's the one article I can remember that I've come across that I've had resolution issues with. For the most part, it seems people usually let the browser size things to fit instead of trying to force extremely wide tables/images.
Anomie⚔17:00, 13 August 2008 (UTC)
Download speed
Something has begun to happen that I haven't experienced before. When I attempt to access Wikipedia, but not other websites, the download process begins, then stops when about a sixth of the info has been downloaded (a guess from looking at the bar at the bottom of my screen). Nothing happens for awhile, sometimes up to a minute, during which time my PC is locked up. Then it suddenly downloads the rest and everything's fine again. What's going on? Other websites download immediately. I have a high speed cable connection.
If your PC is locked up, for up to a minute, why do you think it is "download speed" at fault? Sounds like page rendering or javascript problems. What operating system/version and browser/version do you use? Do you have access to another browser on your same machine to test it? Does this happen with Javascript disabled? Does this happen on other Wikimedia projects, or other language Wikipedias? --
Splarka (
rant)
07:34, 7 August 2008 (UTC)
You're quite right about download speed. It must be something about my setup. I'll check it and return later. I use Windows Vista. -- Fyslee /
talk19:48, 7 August 2008 (UTC)
I recently had my PC's graphic card replaced, and it looks like the technicians must have done something to my Java. I have reinstalled it and now things work fine again. Thanks for the tip! -- Fyslee /
talk04:32, 8 August 2008 (UTC)
So much for that! It only worked fine for about three times and then back to the old long stop and then suddenly go. The Task Manager says "Status: Not responding" during the pause. What else could be wrong? -- Fyslee /
talk06:05, 8 August 2008 (UTC)
What else could be wrong? Try some of the things I (indirectly) suggested above: disable javascript (not java, separate things). Try other projects (
wikt:v:s:), other languages (
fr:de:es:), etc. --
Splarka (
rant)
07:18, 8 August 2008 (UTC)
I doubt JavaScript has anything to do with it. However, as
Googling for "IE7 disable JavaScript" shows, the setting should be under Tools -> Internet Options... -> Security -> Custom Level... -> Scripting, "Disable", hit OK and restart browser. (That's if you're using Internet Explorer 7, which seems like a safe bet.) —
Simetrical (
talk •
contribs)
20:42, 12 August 2008 (UTC)
I use Windows Vista and everything else (including all other websites) works just fine, so it's not the browser. I've never had this problem before with Vista. This is something new and odd. Very irritating. -- Fyslee /
talk19:22, 11 August 2008 (UTC)
That other websites work fine doesn't mean it's not a browser bug. Wikipedia might just be the only website you're going to right now that's triggering the browser bug. If you
try Firefox (free download, should take only a few minutes to try out) and it works, then you know it's partly Internet Explorer's fault. —
Simetrical (
talk •
contribs)
20:42, 12 August 2008 (UTC)
Years ago, my IE installation got mangled in a subtle way that caused it to have problems on certain sites. I was able to get those sites to work again by using Firefox instead. So, like Simetrical says, try Firefox... you might just find you like it better than IE in general, in which case, problem solved, and then some.--
Father Goose (
talk)
02:04, 13 August 2008 (UTC)
No, it would not. Most significant bugs are caused by the interaction of many different aspects of the computer in unexpected ways. It could very well be that the combination of Wikipedia plus Internet Explorer plus something else is causing the issue. In fact I would say it's very likely, but that's a guess. The way you figure out what the problem is is by changing one thing at a time until you narrow down which components are necessary for the problem to occur. If you would be so kind as to do so, we can proceed with debugging. Otherwise I guess I can't help you, but good luck. —
Simetrical (
talk •
contribs)
16:35, 13 August 2008 (UTC)
impersonation using old usernames after a username change
I see that, once you change your username, your old username is open for registration to any random troll that wishes to impersonate you. See
user's call for help and
relevant ANI thread.
/me thinks that, when you change a username, the old name should be automatically be invalidated for registration. See also
WP:BEANS, oooopsie.... --
Enric Naval (
talk)
13:05, 9 August 2008 (UTC)
Just wondering if there would be anything that we could do to the redirect fixer bot to stop things like
this happening? Maybe add some filters or something? ·Add§hore·Talk/
Cont14:09, 10 August 2008 (UTC)
Short answer is "no" - the RedirectFixer user is actually part of the MediaWiki software, and can't be blocked, modified or coerced (:D). However, IIRC the script, when editing pages, looks for the current redirect target, not the old one. So if you revert the pagemove quickly enough, the script will make null edits to all the double redirects, which won't show in the edit history. Alternatively, just by waiting a while, the script will revert itself when it sees that the pagemove has been reversed.
Happy‑
melon16:19, 10 August 2008 (UTC)
Yes, when the pagemoves are reverted, redirectfixer is generally fast enough to re-update redirects before the vandalized target is deleted. If moves are expected (rather than vandalism) to a page where the redirects should stay put (eg, move-archiving), there is a magic word to prevent the redirect updating.
Gimmetrow16:26, 10 August 2008 (UTC)
Eh, what would you put on this redirect-fixer regex blacklist beyond what is already on the page-move/creation regex blacklist? —
CharlotteWebb19:36, 12 August 2008 (UTC)
Hmm good point :P. Maybe have a longer delay before the bot makes the edit? Then before it makes the edit check to see if the move is still valid? ·Add§hore·Talk/
Cont08:26, 13 August 2008 (UTC)
Couldn't the reaction time be reduced? If it didn't react as fast as move vandalism happened, then it wouldn't have to fix the ensuing cascade of echoed vandalism it causes (and neither would anyone else). --
Muna (
talk)
17:02, 24 August 2008 (UTC)
Page deletion revision limit
Is there a way to circumvent the 5000 revision limit on page deletions? The reason I ask is the limit was hit when trying to process a history split & merge of page
Apple.
Keith D (
talk)
11:19, 12 August 2008 (UTC)
No, only the devs can handle deletes larger than that IIRC (same for renaming users with over 200,000 contribs).
Happy‑
melon12:02, 12 August 2008 (UTC)
*An old one that needs special action as there are over 5000 revisions, so I have not tagged articles. Revisions of
Apple 12 June 2003 and earlier are revisions for the dab page at
Apple (disambiguation) and should be moved and merged there. Revisions of
Apple (fruit) 30 June 2003 and earlier (all apart from today's cut & paste attempt) are the early revisions for
Apple and should be merged there.
Keith D (
talk)
11:36, 11 August 2008 (UTC)
I tried to delete page
Apple temporarily to history-split off its first 8 edits (the disambig edits) as the start of this job, and I got a fault page showing "This page has a large edit history, over 5,000 revisions. Deletion of such pages has been restricted to prevent accidental disruption of Wikipedia." in red. Leave it and merely put notes in the affected pages' talk pages to explain the cut-and-pastes? Or what happens?
Anthony Appleyard (
talk)
14:27, 11 August 2008 (UTC)
That is why I listed here as I got the same message. I was hoping that someone here would know if there was a way of fixing with the limit in place.
Keith D (
talk)
15:36, 11 August 2008 (UTC)
I believe that stewards are technically capable of doing this (by granting themselves the bigdelete right). However, since the whole point of the limit is that Bad Things can happen if you violate it, I agree that if anyone does so, it should be the devs.
Algebraist15:00, 12 August 2008 (UTC)
In this case, the best option would probably be to create a talk subpage for each article, then copy and paste the relevant lines from the history into it. Then you could make a
dummy edit to the actual article to link to where the history is copied to.
Mr.Z-man15:41, 12 August 2008 (UTC)
I just was editing
Canada and made a very particular change after reviewing a source. The change might not make nearly as much sense without an edit summary, but I accidentally left it so by hitting the return key as well as the right-hand shift key while typing my edit summary, resulting in a blank edit summary. Would it be possible, from a) a technical perspective and, b) a social perspective (not gameable), to be able to create some sort of short "grace period" where one could correct a stupid mistake in an edit summary? It would certainly be useful, and I'm curious whether it would be feasible in practice. Opinions; suggestions; comments? {{
Nihiltres|talk|
log}}
04:01, 13 August 2008 (UTC)
If this happens then just do a null edit, where you add a space or something and then use the edit summary how you wish. GaryKing (
talk)04:08, 13 August 2008 (UTC)
Note that this is called dummy edit, while null edit is when you save without any changes, just to update the page (if action=purge doesnt help). —
AlexSm17:14, 13 August 2008 (UTC)
Apart from being a textbook example of why we desperately need 2-dimensional referencing, the only problems I can see in that reference list are caused by using square brackets to close templates. There are numerous cases of code like {{cite something|... ... ]] when they should be {{cite something|... ... }}. I think there's also one case where extra sets of curly braces have been included unnecessarily. Have a look through and check that the closing braces of each citation template match the opening braces, and you should be ok.
Happy‑
melon13:24, 13 August 2008 (UTC)
OK, I think I've done them all, there is still one "refname" with no matching ref, hopefully whoever was adding the refs will notice that some changes have been made and spot that one.
DuncanHill (
talk)
13:43, 13 August 2008 (UTC)
I have a HP Scanjet 4850 and I run a Mac G4. My previous scanner had a feature that, when scanning photos from newspapers, would "blend" the print dots so the resulting scanned photo would be smooth rather than a bunch of various sized and shaded dots. I don't see how I can do that with this scanner. I have been frustrated by going to the HP site for help. Does anyone know if there is a way I can achieve this result. By the way, I usually scan into photoshop.
(
Eleeguy (
talk)
15:12, 13 August 2008 (UTC))
It's working fine on IE at the moment. Are you using a different browser? Or it may just be your computer. Lradrama17:29, 13 August 2008 (UTC)
Safari on WinXP, it's speeded up again, but for a time pages such as this one and the refdesks and the admins' noticeboards would not load fully.
DuncanHill (
talk)
19:36, 13 August 2008 (UTC)
Odd duplicated text
I made
this edit just now which, on my browser at least, resulted in a duplication of the words "albums/tours" on
the resulting page. Does anyone else see it, and any ideas why? I've never seen this particular behavior before.
Jgm (
talk)
16:32, 13 August 2008 (UTC)
Just IE6. It seems to be a browser issue: I can cause different words to be duplicated (always at the end and beginning of a line break) by changing the browser width. Weirdly, this talk page is the only one I seem to be able to make it happen on, and then only on my edit (and now the one below it).
Jgm (
talk)
17:47, 13 August 2008 (UTC)
Ok. I know that IE6 has some bugs that cause duplicated characters in some circumstances. This seems to be another variation of it. ---
RockMFR19:29, 13 August 2008 (UTC)
Regex search of DB dump
I want to run a regex search against the wikipedia database to identify a list of article for my bot. Is my only option to download a copy of the whole db (pages-articles.xml.bz2
3.9 GB), decompress it, and perform my simple little search? This seems like an outrageously inefficient use of resources for both me and the wikimedia foundation. Is there no facility that allows me perform the search remotely? Are there users who already have copies of the database and are willing to generate article lists for others?
PlasticupT/
C17:49, 13 August 2008 (UTC)
I asked this question at the Help Desk but received no answer, so I suppose it is a little technical and maybe someone here can help me. Are either of the following possible to obtain:
a URL which adds a preload template to the bottom of an existing page. (Normally the template is ignored if the page exists - even if the page is blank somewhat annoyingly.)
a URL which creates a new section of a page, but pre-fills the name of the new section. (
This URL takes you to a page where you are prompted to enter the name of the new section. I want to specify the name and not prompt for it.)
but each time it never reads more than 51 articles. How can I get it to read more in each tranche? Is this a matter for Bugzilla or a local setting?
To be "green" I would like to be able to switch off my computer when not at it and then read a massive block of up to the last twelve hours of new pages when I come back. But this might require a mechanism whereby I could register myself as a responsible user, not likely to grab masses of bandwidth. —
RHaworth (
Talk |
contribs)
22:24, 13 August 2008 (UTC)
Could someone here please tell me what the purpose of the "Go" button and the checkbox marked "Invert Selection" on my watchlist means? I can't figure out what selection it's inverting, or what the "Go" button's supposed to do. -Jéské(
v^_^vBodging WP edit by edit)06:51, 14 August 2008 (UTC)
Choosing a namespace from the dropdown will filter the returned results, limiting them to that namespace. Checking the invert box will cause all namespaces but that one to show (useful for filtering out particularly spammy namespaces, more useful on recent changes than on most watchlists). The [Go] button simply activates this choice. --
Splarka (
rant)
08:16, 14 August 2008 (UTC)
Hi, the drop-down buttons to select namespaces at e.g. my contributions, what links here, etc have stopped working (I use Safari on WinXP). I recall this happenning before.
DuncanHill (
talk)
11:10, 14 August 2008 (UTC)
Is it possible to use a bot to generate a list of articles with geographical co-ordinates within the Arctic Circle? That would be a good approximation to suitable articles, though there are other definitions of Arctic, and subarctic stuff is sometimes relevant as well. Still, all articles with latitude co-ordinates between 66° 33′ 39″ (or 66.56083°) and 90° North. Is that possible? Where should I ask?
User:Carcharoth/Arctic articles is one place to put the list (or link to it from there).
Carcharoth (
talk)
12:07, 14 August 2008 (UTC)
Edits before 10:45 am EST today do not appear in any user's contributions? Is this happening to anyone else besides me? GaryKing (
talk)18:31, 14 August 2008 (UTC)
Same thing here. You are not going crazy. I'm also getting a "You have new messages" bar but I was the last one to edit my talk page. Ghost in the machine.
KnowledgeOfSelf |
talk18:33, 14 August 2008 (UTC)
I'm guessing the system is lagging, although this particular issue is pretty serious as seeing user contributions is useful in many situations. GaryKing (
talk)18:33, 14 August 2008 (UTC)
(edit conflict x2) I just came back from lunch (1:30 pm CST) and they are gone, from my watchlist AND my contribs. Everything after 09:55 am CST! It was working fine when I left at 12:19 pm CST. (and, as a note, this edit also did not show up in either place...adds a whole new meaning to LAG) --
AnmaFinotera (
talk·contribs)
18:34, 14 August 2008 (UTC)
That's what I thought, Xeno, but it's not the most recent four hours dropping out of the watchlist; the same recent edits are sitting there, even though subsequent edits have been logged and are visible at the article history itself. It's not a moving window, it's a growing window. And it will consume us all...UltraExactZZClaims~
Evidence18:39, 14 August 2008 (UTC)
Did anyone file a bugzilla yet? I just made an account but I'm not too familiar with how to do this. –xeno (
talk)18:42, 14 August 2008 (UTC)
This seems to have occured in the last half-hour or so, as a few issues I was following suddenly lost their contribution histories. (My watchlist also stops at 7:53 AM PDT today.) Any idea what is happening? --Ckatzchatspy18:42, 14 August 2008 (UTC)
Edits are slowly coming back now. I'm getting "Due to high database server lag, changes newer than 13395 seconds might not be shown in this list." and edits from 11:00 am to 11:05 am have now appeared for me. GaryKing (
talk)18:53, 14 August 2008 (UTC)
Ah, so a server crashed and I imagine it has now been replaced which is why changes are coming back. Watchlist and contributions grab changes from multiple pages so they are cached and have a lower priority than page histories I imagine, so they can die as long as page histories live :) GaryKing (
talk)18:58, 14 August 2008 (UTC)
Is still not showing anything I don't like this because I always use my watchlist to update me on changes. Anyone can help that would be great?!
KakiBrz (
talk)
19:58, 14 August 2008 (UTC)
I first noticed this on my talk page yesterday. I have a "top quotes" box transcluded there. It suddenly disappeared.
But the navbar didn't, so I compared them, and saw that the navbar didn't have the boilerplate metadata classes.
So I removed them, and the box reappeared.
But then, I'm now noticing I'm having the same trouble viewing CfD pages. And noticed that they also have the "boilerplate metadata" classes.
I'm guessing something was "changed" somewhere. Can this be fixed?
(And while you're at it, maybe fix the class which used to display hidden boxes when javascript wasn't present which now makes them hidden without means to view except by editing the page. Doczilla's RfA thank you on my talk page is an example. In this case, it happened prior to his posting.)
No. But the backlog there is apparently immense. And this is directly problematic for me right now.
Consider: I look over
WP:AN, and if someone has closed a discussion, that discussion disappears. If there isn't a reference to the "box" nearby, I may not even know that it was ever there. Imagine someone linking to a discussion with the hopes that you will read it as it may have bearing on the current discussion, but when yo ugo there, you only see part or none of it. That may have a bearing on your discernment.
And I have to say that not being able to see
Template:Otheruses (among others, like
Template:Main), as a reader, is driving me nuts.
Have you tried a different browser? I don't see any similarities between
Template:Otheruses,
Template:Main, and class="boilerplate metadata". Sounds like the issue is on your end, else I'd imagine we'd have heard a lot more problems by now. Perhaps your browser has borked CSS handling? --
MZMcBride (
talk)
06:55, 11 August 2008 (UTC)
Does it happen in a different browser? If using Firefox, does it happen with a fresh profile? Does it happen when you're logged out? Try disabling any Gadgets you're using. Does it happen on other Wikimedia sites, like Wiktionary or Wikipedias in other languages? Does it happen on any non-Wikimedia installations of MediaWiki, like Wikia wikis? Does it happen if you disable JavaScript? If you disable CSS? On other computers?
A quick answer to the above: I think the main problem is likely that this computer doesn't run javascript at all. (Including gadgets and whatnot.) And some recent change (I am guessing) requires javascript.
Also, when I remove the classes (in particular the two mentioned above), the templates display fine. That's why I was guessing that they're involved somehow. But it's obviously only a guess on my part. -
jc3718:39, 11 August 2008 (UTC)
So the answers appear to be:
It only happens in IE.
You're using no JavaScript, and haven't tested enabling it.
It doesn't happen on other computers.
Questions I asked (some of which others have asked as well) that are still unanswered:
Does it happen when you're logged out?
Does it happen on other Wikimedia sites, like Wiktionary or Wikipedias in other languages?
Does it happen on any non-Wikimedia installations of MediaWiki, like Wikia wikis?
Does it happen if you disable CSS?
Also:
Does it happen if you enable JavaScript?
What version of IE are you using?
Does it happen if you change your skin in your preferences?
Please systematically answer all of these so that I can try to help you debug the issue properly. And by the way, it's not likely that this has anything to do with the developers. It's most likely some enwiki-specific thing. None of the classes or templates you're having problems with are in the core software, so it's not likely bugs in them are the fault of the developers. —
Simetrical (
talk •
contribs)
20:36, 12 August 2008 (UTC)
Wow. Thank you. A few of these I haven't tried, though I will shortly.
Also, my sincere apologies for anything that I may have said which may be construed to indicate anything untoward towards anyone. -
jc3721:40, 12 August 2008 (UTC)
Ok, attempted. Though I'll admit at times it felt like adventures in wonderland. (Trying to find things not where you're used to finding them, or in other languages : )
So here's what I found. I hope it helps.
Does it happen when you're logged out?
Yes - I checked
WP:CFD (August 4 - closed cfds do not display, merely the headers)),
Template:cfd top (which doesn't display at all),
WP:UBX (seealso shows up under the usercat section, but not main), and the version of my sub-page top-quotes before I removed the classes (calling up the last June version from the page history), which doesn't display at all.
Does it happen on other Wikimedia sites, like Wiktionary or Wikipedias in other languages?
While not logged in (tested with
Template:cfd top), picked at random:
norsk wikipedia - no apprearance.
dansk - appears!
deutsch - appears
latina - appears
espanol - appears
teing viet - no appearance
islenska - no appearance
bahasa indonesia - appears
So far it looks like the wikipedias in the last section on the main page have "no appearance", while those in the first two sections the template appears.
As for sister projects:
Commons - appears
MediaWiki - appears
Meta - does not appear
Wikibooks - appears
Wikinews - appears
Wikiquote - appears
Wikisource - appears
Wikispecies - appears
Wikiversity - appears
Wiktionary - appears
Does it happen on any non-Wikimedia installations of MediaWiki, like Wikia wikis?
I tested this by editing a page, replacing it with the contents of Template:cfd top, and previewing the page (not saving). It displayed at every one I tried (and in those which I had an account, both logged in, and not). While at en.wp, doing the same test, it did not display.
Does it happen if you disable CSS?
AFAICT, I have no personalised style sheets in use.
Does it happen if you enable JavaScript?
Not an option here. Though the other computers I tried all had/have javascript running, which presumably may be why they didn't have the issue?
What version of IE are you using?
Not xp or vista
Does it happen if you change your skin in your preferences?
(Still using template:cfd top):
chick - appears, but the search box buttons didn't work.
Okay, good. Yes, it looks like a broken customization of some kind in Monobook.css, that probably several of the large sites have copied from each other. A couple more clarifications, if you would:
What version of Internet Explorer are you using? To find out, go to Help -> About Internet Explorer, or something similar to that. (To get to the menus you might have to hit Alt first or something, in recent versions.)
You said "not XP or Vista": what version of Windows are you using? 2000, Me, 98, 95, some server version? It will probably say when you start up your computer, or else feature in the interface. If you posted a
screenshot of your computer screen with Internet Explorer open it would answer this question as well as the last.
Why isn't it possible for you to try briefly enabling JavaScript? Is the computer owned by someone else and locked down or something like that?
Note that disabling JS may have something to do with it, but I see no problems with JS disabled on Firefox 3, and I'm pretty sure others would have noticed if disabling JS on other browsers did it. I'm not sure, but right now I'm guessing from your answers that you may be using a very old version of Windows and/or Internet Explorer, or otherwise a very strange setup, which could be why no one else has noticed a problem. —
Simetrical (
talk •
contribs)
16:30, 13 August 2008 (UTC)
I once copied a custom skin and overwrote my monobook.css file;
Today i wanted to revert my skin to the original monobook, and i copied the default css found on wikimedia commons; i also copied the default monobook.js to my monobook.js page and added a couple of custom scripts for editing tasks.
Problem
I purged my firefox cache and switched back to monobook from my preferences page, but it keeps loading the custom css i had before.
I tried this many times but to no avail.
Is there a class of table that sorts itself in reverse oreder by default when clicked on? With the "sortable" class you have to click on the little icon twice in order for it to sort in reverse order. Thanks.
SharkD (
talk)
22:21, 14 August 2008 (UTC)
The
particular table I'm working on uses a binary system where cells with content in them are of interest and blank cells (the vast majority) are not. Blank cells unfortunately always get sorted to the top. Having the reader only need to click once means there's fewer hoops for him/her to jump through (especially since there are about a half dozen or so of such columns).
SharkD (
talk)
02:22, 15 August 2008 (UTC)
That would work, but would complicate the table and use considerably more parser directives. Where can I find a page that describes the JavaScript files used on Wikipedia? I found the one that describes
CSS files as well as the page that describes
user scripts. I've worked with sorting routines in JavaScript before and don't recall them using much overhead in simple cases such as these. I don't think an addition/modification would be too big of a problem in this case.
SharkD (
talk)
03:05, 15 August 2008 (UTC)
I found the page,
here. This particular routine is a bit more complicated than I expected, but I'll take a look and see what I can come up with.
SharkD (
talk)
03:20, 15 August 2008 (UTC)
One more thing: is there a way to sandbox the changes to this script? I'm unfamiliar with user scripts. Can I install a user script that overrides the relevant bits of "wikibits.js"? Thanks!
SharkD (
talk)
03:27, 15 August 2008 (UTC)
Thanks, I'll give it a shot. Also: is there any support for changing the script so that a third click causes the template to return to its initial, unsorted state? Currently, once you click the icon there's no way to view the template in its original state without refreshing the page. Changing the script in this manner should relieve some stress from the servers, as the all the work done is client-side instead of on the server.
SharkD (
talk)
19:07, 15 August 2008 (UTC)
Yes, that is the point. However, I also want to include the span tags in the template so that you need only wrap the text in the template. I'm just checking if anyone has already created such a template before I create my own.
SharkD (
talk)
19:11, 15 August 2008 (UTC)
Work lost when database is locked
Is there any way to preserve one's work when the dreaded "Database locked" page pops up? Are there two types of messages, one where one's work is preserved and one where it isn't?
SharkD (
talk)
02:17, 15 August 2008 (UTC)
Is there a separate case (i.e., a different message) where the content is preserved, or was edit retention disabled recently? I'm confused.
SharkD (
talk)
02:45, 15 August 2008 (UTC)
Neither. I remember (incorrectly?) that in some instances your changes are (were?) retained on the page where the error messages are displayed. They're not saved to the database, but they are presented to you in a text field so you can either try to continue editing or else back up your changes. Maybe I am imagining things... ?
SharkD (
talk)
02:58, 15 August 2008 (UTC)
Also, I am curious. Sometimes in IE you can refresh or go back to a page and all form fields will remain filled. What is different in this case that keeps it from happening?
SharkD (
talk)
03:11, 15 August 2008 (UTC)
IE is notoriously unreliable. If there is not a pressing reason you cannot install Firefox (i.e. locked down company CPUs), I would highly suggest you upgrade. I used to be an IE guy myself...but after using Firefox religiously for the past year or two, I could never go back. –xeno (
talk)03:47, 15 August 2008 (UTC)
Yes, FireFox is much more relible; I have been using it for the last 2 years. Recently, I was visiting a friend and used his computer to edit wiki for several hours. The result—IE7 crashed couple of times (and offered to send a report to MS, which I politely declined). In one such crash I lost rather a long piece of code in the template, which I was working on. So IE7 is not for me.
Ruslik (
talk)
05:53, 15 August 2008 (UTC)
Actually, IE7 on Vista runs in limited-privilege mode with no direct filesystem access, so it might be more secure at this point. Although I'd still recommend Firefox for its superior features (or Opera, etc.). —
Simetrical (
talk •
contribs)
14:19, 15 August 2008 (UTC)
I use Avant, which is a wrapper for IE. It has more features by default (e.g., minus plugins) than FireFox, and is a bit more streamlined as well.
SharkD (
talk)
19:28, 15 August 2008 (UTC)
Normally, when there's a database lock and you make an edit, you get the form back, so you can just try to save again, or copy it into a text editor to save it, etc. It basically translates the submit into a preview. If you ever see a database lock error and you don't get the text you submitted back so you can copy or resubmit it, that's a bug, please say where it happens. However, others are correct that in other browsers than IE this isn't a problem: if you hit "Back" your text is still there, even if the software doesn't give it back to you for resubmission. —
Simetrical (
talk •
contribs)
14:19, 15 August 2008 (UTC)
Hmm, maybe a "try again" button could be added to the page which appears when your edit can't be saved due to db-lock? —
CharlotteWebb15:05, 15 August 2008 (UTC)
Thanks. That's the information I wanted. One other thing: are database "locks" scheduled in any way? What I mean is, is there a way to predict when they occur beforehand?
SharkD (
talk)
19:04, 15 August 2008 (UTC)
No, they are not planned. They are the result of temporary spikes in the work the database is asked to do. — Carl (
CBM ·
talk)
19:33, 15 August 2008 (UTC)
I was editing the page on Prior restraint and when I pressed save, the page had a weird color and cyrlic (russian) letters with the words. "I am the Zodiac you cannot stop me". With a weird picture diagram. This is of course some hacker trying to play off on the Zodiac killer but the main news here is that Wikipedia has been hacked! It is no longer safe. Please turn on your restrictions. —Preceding
unsigned comment added by
76.181.52.42 (
talk)
18:17, 15 August 2008 (UTC)
Yes, I was being mildly sarcastic when I suggested it. It is bolting the stable door after the horse has bolted. Not one template has been hit twice by the recent "Zodiac" vandal.
Woody (
talk)
20:51, 15 August 2008 (UTC)
I took a quick look at the number of transclusions and figured that it'd be a good idea to semi-protect it. If you disagree, you're more than welcome to unprotect it; I really don't care one way or the other. :)
EVula//
talk //
☯ //20:55, 15 August 2008 (UTC)
Not bothered enough to overturn you, just commenting on it, I remember someone getting a bit bothered about it a few days ago on one of the AN boards. I am sure the glorious new task force
Wikipedia:Page protection patrol will be along if anyone really objects! ;) Just seems we should carry on blocking, reverting and ignoring (and calming down newbies who think we have been hacked by zombies ;) Thanks and regards.
Woody (
talk)
21:01, 15 August 2008 (UTC)
Has no template been hit twice because of a conscious decision by the vandal or because they are almost always protected shortly after the vandalism is reverted?
Mr.Z-man21:06, 15 August 2008 (UTC)
Yeah, we've been bombarded with complaints about images and them not resizing or something (Help Desk, here, and AN/I both have threads)
Calvin 1998(
t-
c)05:23, 16 August 2008 (UTC)
This is a good idea, however I can think of a few issues with it. It seems to me that this template would require a lot of if (or case) statements in order to match up the inputed country code with the correct output text and image, plus there would need to be error handling in case of an incorrect code. This would create a very complicated template that only a few people would be able to understand and be able to modify. If a new country was created, even an extremely minor error in adding the appropriate code could really mess up the template. Although a unified template would be much easier to use, the complicated code that's required would be a problem. I would suggest concentrating instead on standardizing both the design of the country templates and the name of the template. --
Imperator3733 (
talk)
02:47, 16 August 2008 (UTC)
This could eaisly and simply be created by using switch statements instead of if statements. Switch statements also have error handling. Support. —Atyndall citationneeded09:15, 16 August 2008 (UTC)
Switch statements (or large numbers of conditionals) are very expensive. Use a template that redirects to a large directory of sub-pages instead of using parser directives. The {{getalias}} template (or some other template based off of
Country data) should suit your needs.
SharkD (
talk)
09:22, 16 August 2008 (UTC)
Page existence confirmation
I'm not sure where to ask questions about bots and scripts. If this is the wrong place, please direct me to the right one. I'm trying to write a (cgi) script that checks the existence of wikipedia pages in more than one language wikipedia. en.wikipedia.org, sv.wikipedia.org etc. I construct the urls with no problem: en.wikipedia.org/wiki/name etc, but when inspecting the returned content I have problems finding any simple criteria that tells if this actually is the wanted page, or just the page telling that the page doesn't exist. Every national wikipedia has different layout and phrasing, even in the title. But there are supposed to be global bots out there, so this problem must have been solved. Anyone?
Akebrett (
talk)
20:35, 15 August 2008 (UTC)
This webpage has an interesting scheme for presenting the reader with information on music notation (see "Interactive Notes"). It shows a staff and the individual notes play a sound when clicked. How hard would it be to implement something like this in Wikipedia?
SharkD (
talk)
10:33, 16 August 2008 (UTC)
I am trying to create a new account over at
mw:, but it doesn't let me do it as it says that the username I try to use for that account ("It Is Me Here") has already been taken by the Unified Login system (which is perfectly true). However, going from
w:en: straight to
mw: doesn't work either (i.e. it doesn't automatically register me) and so I am rather confused as to how I can create an account over at
mw: and retain my user name. NB I tried to use the same email address there as I use here, but to no avail. Any help would be greatly appreciated.
It Is Me Here (
talk)
12:12, 16 August 2008 (UTC)
Just now, when browsing image pages it seems that what was previously a large template with "This image is located at Commons" has turned into a small amount of inline text. I preferred the large template. Have I accidentally done something wrong to my preferences or has someone recently changed that template?
It Is Me Here (
talk)
13:14, 16 August 2008 (UTC)
From
Special:AllMessages, it looks like you're seeing the default values of sharedupload and shareduploadwiki-desc. It should be possible to edit the en-gb versions, but I don't know how.
Algebraist17:48, 16 August 2008 (UTC)
Any chance someone could update
this tool? I think it would be very easy to do - it looks like it only needs the addition of a couple of lines - except that I don't know how to upload a page onto the web.
The tool counts page hits for any wiki page, but its creator,
User:Henrik, has not edited since early July and it cannot currently show edits later than June. Lots of users would like to know how many page hits their articles are getting!
Gatoclass (
talk)
15:33, 16 August 2008 (UTC)
Can someone familiar with the system software updates made back in January update and correct the old behaviors documented (pre-bug-fix, as it were) in
On going practices on the commons suggest these are now very inaccurate. Hard category redirects are being used all the time now. // FrankB16:25, 16 August 2008 (UTC)
This needs much more discussion. I know you prefer to use hard redirects for category redirects. But that doesn't work well for a number of reasons, even though MediaWiki has been partly updated to handle hard redirects for category pages. Among other things MediaWiki doesn't make the pages in the redirected category show up in the category the redirect points too. This means that pages in the redirected category gets "invisible" in the category system, since if you go to their category you don't see them since you get automatically redirected to the other category. Unless you manually go back by clicking the "Redirected from" message at the top of the new category. So we use soft redirects instead by placing the {{category redirect}} template on the redirected category page. That template also helps a bot find the redirected categories so it can automatically move the pages to the new categories.
I suggest you bring this up for discuss at
Template talk:Category redirect and try to change the current consensus. (Instead of trying to force your will on the rest of us by leaving rude messages on our talk pages like you did with me some day ago.)
How do I switch my default embedded .ogg player in Firefox?
It keeps reverting to Quicktime, but I find that VLC media player works much better. How can I get rid of the Quicktime plugin (not get rid of it per se, because I need it for iTunes to work; I just want to disable it during Firefox sessions) and use VLC for all .ogg files?--
The Fat Man Who Never Came Back (
talk)
17:19, 16 August 2008 (UTC)
I think you're a couple steps ahead of me. I don't know how to get VLC player to pop up in the first place, when I click on the play icon for .ogg files on Wikipedia. Where is this option controlled (the option that decides to use Quicktime)?--
The Fat Man Who Never Came Back (
talk)
17:33, 16 August 2008 (UTC)
(click the play button for any
sound file, and a "more..." link should appear. click that for the options). I used to have a VLC option under a Quicktime option. However, both those options have disappeared for me, in the last few weeks, to be replaced by the solitary "mplayerplug-in" option. Most odd; I don't recall messing about with plugins or preferences recently... --
Quiddity20:28, 16 August 2008 (UTC)
The fact that it's changed isn't particularly concerning, I did a whole lot of work on it. But if you're saying it's incorrect now, and you don't have
this Mozilla plugin, then I'd very much like to see the contents of your about:config (type it into the location bar). The plugin in question is one of the many that pretends to be QuickTime, but fails when you actually treat it as QuickTime. That's why you would have previously had a non-working QuickTime option. --
Tim Starling (
talk)
04:43, 17 August 2008 (UTC)
I couldn't install the mplayer plugin (with the msg ""mplayerplug-in" will not be installed because it does not provide secure updates". Will not be?? Who's computator is this anyway???) I looked but couldn't find a workaround? Is there one or a different source? Mozdev doesn't seem to have it. Thanks.
Saintrain (
talk)
15:14, 17 August 2008 (UTC)
Tim: Sorry if I confused things with my tangential comment. I do have mplayerplug-in working correctly. VLC is not mentioned in my about:plugins (or about:config) but I swear it was an option for oggs a few weeks ago! (No matter, my ubuntu is overdue for a version-upgrade anyway.)
Saintrain: Just in case you aren't aware, I should point out that mplayerplug-in is for unix/linux only (whereas vlc is cross-platform (and
mplayer itself is also cross-platform)). Hope that helps. --
Quiddity17:07, 17 August 2008 (UTC)
Are there any plans of implementing support for the colgroup element in wiki markup? Currently, it's only supported if the table is constructed explicitely from HTML elements.
SharkD (
talk)
00:40, 17 August 2008 (UTC)
See
mediazilla:986. If I understand correctly, if colgroup was supported by all browsers, developers would premit it in the wiki markup a long time ago. —
AlexSm03:58, 17 August 2008 (UTC)
I would like to propose that we move the
Main Page to
Portal:Wikipedia. Currently, the main page is in the article namespace, which causes lots and lots of little problems. Moving it to the Portal namespace instead would offer a number of benefits, including:
The top-left tab would read "portal" instead of "article".
People who want to make copies of Wikipedia, such as people who provide computers to schools in Africa that can't get Internet access, would have an easier time separating actual articles from project content which they don't want to copy. Because the content of the main page changes dynamically from day to day, it would take quite a bit of work to make the main page work and keep working on an offline copy of Wikipedia. Thus, since the main page won't actually work by default, it's probably best to exclude it from copies of Wikipedia article content by default.
The "cite this page" link in the sidebar would be hidden from screen readers and text-only browsers, and the
sitewide CSS would no longer have to contain a special declaration to hide it.
The article count shown at
Special:Statistics would be accurate instead of being 1 higher than the actual number of articles on Wikipedia. {{NUMBEROFARTICLES}} would also be accurate instead of being off by one.
Statistics about Wikipedia articles would be more accurate and not slightly skewed by statistics about the main page that are likely to get mixed in.
It would be generally easier to write bots and other automated scripts because developers would not have to worry about having to write special code for the main page, ever.
Here's how the proposal would actually be carried out:
After the links have been transitioned and everyone is comfortable with
Portal:Wikipedia (perhaps a month or two after step 1),
Main Page would be changed to a simple redirect to
Portal:Wikipedia.
Please remember that the main page appears as simply "Wikipedia, the free encyclopedia", so the only thing that will change for readers is the URL, and in any case they will be able to continue using
http://en.wikipedia.org/wiki/Main_Page instead of
http://en.wikipedia.org/wiki/Portal:Wikipedia as long as they like. In the long run (think years), users will access
Main Page less and less until it is used so infrequently that we can delete it altogether.
Interestingly, the German Wikipedia has already done something similar. They moved their main page to the Wikipedia namespace (
de:Wikipedia:Hauptseite), and it is working fine for them.
Some of them may seem ridiculous, but only if you ignore the real issue and focus on the tinier secondary effects. The Main page is not an article, it does not belong in the article space. End of story. --
NickPenguin(
contribs)02:07, 14 August 2008 (UTC)
Suggestion - I've been a web master since 1995 and ever since I first saw the Wikipedia main page in 2004 I have thought the name is wrong. There is a long standing tradition on the world wide web what the main page of a web site should be. Most of the major organisations on the Internet including the web and Internet standardisation organisations follow that tradition. The main page of the English Wikipedia should of course be http://en.wikipedia.org/ , nothing else. And for database technical purposes such as what name magic words return inside templates and so on the name can be for instance
Portal:Wikipedia or whatever people prefer. Thus making the talk page be
Portal talk:Wikipedia. --
David Göthberg (
talk)
07:33, 14 August 2008 (UTC)
I'd like that too, but I don't think the wiki software is designed to do that. And like you said, moving the page to a different namespace doesn't prevent us from shortening the URL in the future if that becomes possible. —
Remember the dot(
talk)01:40, 15 August 2008 (UTC)
Comment You are conflating two proposals here. The first proposal is to move main page to the Portal namespace, and the second—to name it Portal:Wikipedia. The second should be discussed only if the first is supported.
Ruslik (
talk)
12:54, 14 August 2008 (UTC)
In the last discussion, several alternatives were proposed and the one that was most supported by far was
Portal:Wikipedia. I'm trying to consolidate the proposal down to focus on this idea, but if you have a better one that hasn't been discussed before then by all means propose it. —
Remember the dot(
talk)01:40, 15 August 2008 (UTC)
Support There are a large number of advantages with this proposal and no real disadvantages. The
Main Page is not an article, therefore it should not be in article space! It belongs in portal space. As mentioned, the only thing that would change to most users is the url, and the old url would still work! As far as I know, every single problem with this suggestion has already been addressed. --
Imperator3733 (
talk)
02:29, 16 August 2008 (UTC)
Oppose I oppose the move and the proposed name. Main Page is not in article namespace, and indeeds predates namespaces. Rather, the software mistakenly acts like it is an article. We should correct the software rather than a perfectly adequate page name. Nor is the name "wrong" just because other websites choose differently.
Superm401 -
Talk07:00, 16 August 2008 (UTC)
Why do you say that
Main Page is not in article space? Pages in article space do not have prefixes, and neither does the
Main Page. Also, take a look at
this page.
Main Pageis in article space. The devs are not going to change the software to accomodate this, especially since everything can be fixed by moving the page. --
Imperator3733 (
talk)
19:27, 16 August 2008 (UTC)
It is not a bug, it is how HTML works. Just don't float images to the left if you need to indent with <blockquote> or <dl>/<dd>/<dt>. Perhaps
this will illustrate the problem. The floating object pushes the text over, but not the text parent object, which itself is still perfectly indented. --
Splarka (
rant)
08:21, 15 August 2008 (UTC)
I can see how for many applications that would be the desired effect. Is there no way around it though? Some way to indicate to HTML that you want the level indented from any floating object. Seems strange that it'd be simply impossible to use a blockquote and a left-aligned image. --
JayHenry (
talk)
13:17, 15 August 2008 (UTC)
It's just how margins and floats work in CSS. Floats compress only line boxes, not blocks, and the margin is part of the block, not the line box. It looks even uglier with borders:
This is a div with a border and background that overlaps with floated images.
If you use an inline-level element like a span, it sort of works, but only for the first line:
This is a span with a left and right margin. The margin doesn't overlap the images, but it applies only to the start and end of the text, before the text is divided into lines. Therefore the left margin only applies on the first line and the right margin only on the last, which is still undesired.
You could trick the blockquote into starting a new
block formatting context, which is supposed to force it to not overlap the float.
This is a div with a border and background that might not overlap with floated images.
Of course, that depends on browsers getting that right. FF3 does, but I would not be surprised if IE doesn't. There also may be other side effects of forcing the new context, e.g. in the above text that would normally overflow outside the div (if any) will be hidden instead.
Anomie⚔16:36, 15 August 2008 (UTC)
Floating content takes it out of the flow of the document. You need to clear the float. Using the example(s) above:
This is a div with a border and background that doesn't overlap with the floated image.
EVula, I can't figure out how to make this work. How can I use {{-}} to place a blockquote to the right of a left-aligned image? Everything I try creates a giant white space, and places the blockquote below the image, not beside it. --
JayHenry (
t)
06:19, 16 August 2008 (UTC)
Oh, I think I got it! A left aligned image, such as this one, seems to bust the <blockquote> tags.
But if you place <div style="overflow:hidden"> just before the blockquote, and a </div> after the blockquote, it appears to work correctly.
So my question becomes... does this cause formatting problems in anyone's browsers? Are there potential issues that come from using the overflow:hidden tag? --
JayHenry (
talk)
15:27, 16 August 2008 (UTC)
I've had bad experiences with overflow:hidden before on older browsers like IE/Mac. It also, of course, affects what happens if there's actual overflow, in a way you don't necessarily want. Compare:
This is a blockquote with a really long word and no overflow: hidden. xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx
This is a blockquote with a really long word and no overflow: hidden. xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx
WOW! did some one of u think about winning the best talk page contribution of the year troophy before ?
Just kidding. i liked seeing the way of discussion, and how simple, and rich in the same time it is, than the point of discussion it self. keep on the good work wikipedians :)
One last pharaoh (
talk)
22:34, 18 August 2008 (UTC)
JS syntax highlighting RegExp error
In JavaScript, if a regexp has a literal quotation mark, ("), the text is highlighted as if the " was a string boundary:
vars="\"abcd\"abcd\""//creates string: "abcd"abcd" (the "s ar literal)//remove all "s with the regxp: /"/gs.replace(/\"/g,"")//This code is supposed to be green, but is blue, as the literal quotation mark in the regxp is treated as a string boundaryvarcolorRectifier="\""//To correct the syntax highlighting, you have to add a dummy string containing a literal "
—Preceding
unsigned comment added by [[User:{{{1}}}|{{{1}}}]] ([[User talk:{{{1}}}|talk]] • [[Special:Contributions/{{{1}}}|contribs]])
But the code still works properly, no? Actually if this is affecting the appearance of an article about javascript it should be fixed, and syntax coloration should probably be disabled in the affected articles until then, as the wrong colors would be more confusing than none at all.
If this is not affecting a particular article I wouldn't worry. You can escape the quotation mark as \u0022 in your monobook or whatever if the colors bother you, e.g. s.replace(/\u0022/g,"") (\u0027 for apostrophe/single-quote) but this could confuse the average reader if used in an article, unless the subject is
obfuscated code, or something. —
CharlotteWebb15:00, 15 August 2008 (UTC)
this is a very common syntax coloring error; I see it in most apps I use that use syntax coloring. it has no effect on javascript execution itself. where did you see this? --
Ludwigs218:39, 15 August 2008 (UTC)
I wouldn't call this a RegExp error. It occurs anywhere double-quotes are escaped, not just regular expressions.
SharkD (
talk)
19:18, 15 August 2008 (UTC)
Nevermind. I see in the example that it works properly in regular strings, so it is in fact a RegExp issue.
SharkD (
talk)
19:24, 15 August 2008 (UTC)
In practice, priority will be given to function issues rather than appearance issues. In practice, most people don't care what their monobook looks like, or even understand half of it, as long as it works. I'm more concerned about readers. If there are articles about javascript, maybe on wikipedia, maybe on wiki-books or some other project, this would be the best solution from now until the syntax-highlight extension is fixed. —
CharlotteWebb22:45, 17 August 2008 (UTC)
How many times on average do you have to refresh a page for it to load, or click the "Save" button to save your edits? For me it just seems to be getting worse and worse!
SharkD (
talk)
08:00, 18 August 2008 (UTC)
On average, once. It's a fair few weeks since I've had any problems. Are you sure it's not your internet connection?
TalkIslander08:24, 18 August 2008 (UTC)
I have found Wikipedia to be kind of balky this weekend, not responding for several minutes at a time, where normally it's been extremely responsive.--
Father Goose (
talk)
09:48, 18 August 2008 (UTC)
How do I determine whether it's my internet connection? This seems to happen only on Wikipedia. It seems that every third or fourth time I have to refresh or resubmit a page or edit. To be honest, though, I'm not as sensitive to this sort of thing on other sites, given that I don't submit any content to those sites.
SharkD (
talk)
21:38, 18 August 2008 (UTC)
Usually once, but in the last week or two the "database locked" keeps popping up at least two times every day and I had never ever seen that notice before.
Admiral Norton(
talk)13:25, 18 August 2008 (UTC)
This page has been viewed ? times last month
As there is a first full month analysis of the wiki's page hits (counted by the squid server), see also
THEwikiStics (full dump available), I wanted to ask if there is an interest to have page view counts included in the page footer again? "This page has been viewed ? times last month" or similar could be added to the wiki footer again by using a bot that puts figures into subpages of a template or so (includable/linkable with /{{PAGENAME}}; the only question is, whether this is allowed in
MediaWiki:Lastmodifiedat by the devs)? --
Melancholie (
talk)
21:56, 15 August 2008 (UTC)
@SharkD: Lower level? @Davidgothberg: Yes, I know. But that tool stopped working; I am considering to continue that tool (as the original author seems to be not available); just thought about integrating stats into pages (like it once had been). --- Best regards,
Melancholie (
talk)
13:03, 16 August 2008 (UTC)
I mean, usually this sort of information is tracked on the server. Since this information is also supposed to be displayed on the server, there's no real need to use a middle-man program that takes the information off the server and puts it back on. The server software itself could put the tracking data into a template, or save it as a system variable (e.g., like {{PAGENAME}}).
SharkD (
talk)
22:24, 16 August 2008 (UTC)
(unindent) If the site has an API, it could be added with JavaScript (perhaps in a Gadget?). I know Henrik's stats (stats.grok.se) have an API.... --
MZMcBride (
talk)
22:49, 16 August 2008 (UTC)
It's not exactly disabled for performance reasons. It's disabled because it's completely inaccurate. The overwhelming majority of page views are cached by Squid and never reach MediaWiki, and those would be ignored in the count. Any page view count needs to be kept by Squid, not MediaWiki. We have such a count now. It just needs to be integrated in some fashion with the MediaWiki interface. —
Simetrical (
talk •
contribs)
15:30, 19 August 2008 (UTC)
Self-bolding link with number sign
This is kind of difficult to explain, but here it goes. Normally, when an article contains a link to the name of that article, the link will show up as bolded (for example, when I type [[Wikipedia:Village pump (technical)]], it will show up as
Wikipedia:Village pump (technical)). However, when that link has a number sign (#) in it, it will not show up as bolded (i.e.
Wikipedia:Village pump (technical)#Self-bolding link with number sign). I can understand why this doesn't "bold itself", as there are several Skip to table of contents links on talk pages with number sign links. However, links with number signs are also used in navigational templates (like in
Template:Green Day, where the "Working Class Hero" link leads you to
Working Class Hero#Green Day cover.) However, on the "Working Class Hero" article, the link to "Working Class Hero on the Green Day template will not "bold itself" because there is a number sign in the link. I think that there should be a way to get a link like that to "bold itself", because on navigational templates the bolded link is an easy way to tell where the article you're on is (as compared to the other articles; for example, on the Green Day template you can easily tell where "Working Class Hero" is chronologically compared to other singles). Is there a way to do such a thing?
XnuxtheEchidna02:35, 16 August 2008 (UTC)
It would be logical only when and if you were currently viewing the section being linked to. It would also be unworkable. I think the current method is fine.
SharkD (
talk)
05:56, 16 August 2008 (UTC)
As per the discussion in
#Reverse-sortable table?, I've made modifications to the table sorting code in "wikibits.js" in order to allow table columns to be sorted in reverse by default. I found that the code to make this work already existed in the source code; it was simply never activated.
Here are are the changes I propose. You can see it in action,
here, if you copy the code to your "
monobook.js". The way I configured the script is to sort a column in reverse order by default if the first cell in the column (typically a table header) has a class name of "sortreverse". If the header cell does not have this class name, the column is sorted as per the usual, ascending order.
I also fixed a minor bug (or deficiency), changed/expanded the ALT text for the little icons (the character being displayed was not supported in my version of IE) and added comments to some routines that were missing them. I also found a potential issue with the code I would like to draw your attention to: In the "ts_resortTable" function a regular expression is used to trim leading and trailing spaces before processing the cell contents. However, the regular expression searches for and replaces the space character with an hexidecimal ASCII code of A0 (equal to 160 in decimal), but this does not correspond to the more commonly-typed space character with a hexidecimal code of 20 (equal to 32 in decimal) that is typed using the spacebar. I wasn't sure if this was intentional or not, but I'd appreciate it if people were to take a look and make a judgement on the issue.SharkD (
talk)
04:02, 16 August 2008 (UTC)
If you've been following along but haven't responded, there were a few issues I missed regarding the icon tooltips. I've resolved those, so please update your monobook.
SharkD (
talk)
05:28, 16 August 2008 (UTC)
I think you're misunderstanding something about the regex. The line is itm = itm.replace(/^[\s\xa0]+/, "").replace(/[\s\xa0]+$/, ""); — that replaces sequences of whitespace (including normal space, \x20) or nonbreaking space (\xa0) at the beginning or end of the string with nothing (that is, it deletes them). The \s includes normal spaces, tabs, carriage returns, and line feeds (and possibly a few other weird things, like form feeds and vertical tabs). So it should be fine. —
KCinDC (
talk)
05:39, 16 August 2008 (UTC)
I also made a proposal earlier to make it so that every third click on the sorting icon would return the table to its original order. I haven't worked on this, though. It requires more substantial changes.
SharkD (
talk)
06:06, 16 August 2008 (UTC)
It would be nice if you posted a diff of all the changes, or (even better) highlighted all the changes with some comments. By the way, I also had several proposals about this code, see
User:Alex Smotrov/mw/sortable; I was supposed to remind Simetrical to look at it but then I somehow forgot. —
AlexSm04:13, 17 August 2008 (UTC)
Here is a diff of the two versions of the table sorting code.
Here is a diff of the entire file, including other changes/optimizations not having anything to do with table sorting.
SharkD (
talk)
05:23, 17 August 2008 (UTC)
What is the best way to submit patches if there are several bugs/issues to report? Should the patches be cumulative? Will whomever reviews the patch be confused if a patch requires a previous patch in order to work? Or, should each patch be built off the current, unpatched state and the sorting/merging of the patches left up to whoever implements the patch(es)?
SharkD (
talk)
21:44, 18 August 2008 (UTC)
I replied on my talk page, where you asked substantially the same question. Basically, if they change different parts of the same file, then submit them separately and they should apply fine. If one patch requires a previous patch, you can either submit them as a patch set on the same bug, or just merge them into one patch. —
Simetrical (
talk •
contribs)
15:31, 19 August 2008 (UTC)
MediaWiki not parsing HTML tags in MediaWiki messages
I think this should be reported to MediaWiki Bugzilla or so, but I don't know how that works so here goes:
I and Oxymoron83 have discovered that MediaWiki does not properly parse some of the MediaWiki messages like
MediaWiki:Anontalkpagetext and
MediaWiki:Sp-contributions-footer-anon. It is likely this affects other such MediaWiki messages but we have not tested that yet. If such messages contain HTML wikimarkup <br> tags, then they are not properly parsed and converted to XHTML <br /> tags. That is, the HTML pages are rendered with <br> tags.
Provided that no one has changed the code in
MediaWiki:Sp-contributions-footer-anon: Take a look at any anon IP user's contribution page, for instance
Special:Contributions/65.103.35.149. See the
MediaWiki:Sp-contributions-footer-anon box at the bottom of the page. (The box with the text "This is the contributions page for an anonymous user".) Then view the source code of the rendered HTML page, see that between the box's text and all the WHOIS links in the box there is a <br> tag. That tag should have been converted to a XHTML <br /> tag before rendering the page.
I have checked both an empty anon IP page, an empty anon page in edit mode, and one with messages on. The
MediaWiki:Anontalkpagetext is shown in all those cases. And in all those cases the <br> tags were unparsed (not converted to XHTML <br /> tags). Also, if we use a HTML table in that message and miss to insert a </td> tag then it doesn't get parsed and fixed like it is in articles and other pages.
Wikimarkup tags such as '''bold''' seems to be parsed and converted as usual.
Note: I have now edited
MediaWiki:Anontalkpagetext to use XHTML tags so it renders right, since otherwise Twinkle did get problems to leave messages on anon talk pages. But I have left the "bugs" in
MediaWiki:Sp-contributions-footer-anon for demonstration, since that should not cause any problems anywhere. (All modern web browsers can handle such old school HTML tags.)
I'm pretty sure this is intentional.
HTMLTidy is not run on everything. The assumption is that admins should be able to produce clean code :) ---
RockMFR22:30, 16 August 2008 (UTC)
RockMFR: As a Wikipedia editor and admin I assumed I should use (HTML) wikimarkup both in articles and in MediaWiki messages. And that MediaWiki then converts that to whatever is the current standard for web page rendering. (Today (2008) that happens to be "XHTML 1.0 Transitional".) Ah well.
Are there important reasons why that's the case, or is it just the usual mottley collection of pages growing organically over five years? I mean, I can see how some messages would be plaintext and some encoded, but what's the reasoning behind some being wikitext, some HTML, and some XHTML? Would it be possible to put some identifying mark on each message to say what syntax was accepted (presumably that message would, perversely, be contained in a MediaWiki: page :D)?
Happy‑
melon21:06, 18 August 2008 (UTC)
(X)HTML messages have been deprecated for some time now, but they're much cheaper to use than wikitext messages, computationally, since they don't need to be parsed. A list right now is hard to compile because the format of the message is decided by the caller, not in any central location. Theoretically (and this does sometimes happen) a message can be used as wikitext in one place in the code, fully escaped text in another, escaped except for entities in another, etc. Of course this is a bug and it's fixed when anyone notices it, but it illustrates the current problem that we don't know what format a message is used in without grepping the code. —
Simetrical (
talk •
contribs)
15:33, 19 August 2008 (UTC)
Trivia and pop culture as randomized infoboxes
Trivia and popular culture references creep into thousands of articles on Wikipedia. These sections are (contrary to style guidelines) often bulletted lists of fictional references and Internet memes. While it's not terribly useful to have these sections in such a form, neither is it valuable on a low-edit article to turn these into well-cited prose only to have several editors come along over the following months and add in successive sentence fragments describing even more of the same.
Instead, I'd like to suggest a model of encouraging the default behavior while also improving the article. This might require an extension to MediaWiki to implement, but what I'd think would work best is a template of the form:
{{trivia list | name=Avocados in popular culture | display=3 | content=
* The video game [[Giant Pit]] (1995) featured an avocado warrior princess with a mayo gun.
* ...
* ...
* ...
* ...
}}
The displayed text would be an infobox containing the "name" as heading and 3 out of the 5 listed items, selected at random (probably at rendering time by JavaScript, though a static version could be provided within a noscript tag). In this way trivia or popular culture sections could be reduced to side-bars in the spirit of "Did you know" or the like, and would not take up large amounts of article space even with hundreds of items.
One of the goals is certainly to reduce the imposition that these sections make on an article, but similarly, this gives editors the ability to communicate less weighty elements of the topic at hand which they might find most compelling. -
Miskaton (
talk)
19:33, 18 August 2008 (UTC)
Any chance you could do an example on say a sub-userpage or something so I can see what it looks like, thanks. Your idea looks promising so far.
Deamon138 (
talk)
23:52, 18 August 2008 (UTC)
It seems like that would just make it more likely for errors in the list to remain uncorrected (since they'd only be displayed some of the time). And it would encourage people to keep reloading the page to see the random items change. Plus, people who are interested in the trivia wouldn't see it all. —
KCinDC (
talk)
00:02, 19 August 2008 (UTC)
As a founder of
WikiProject Popular Culture, I am sensitive to the problems associated with these sections. The real solution is to address them editorially, not technically. As yet we don't have a solid or consistent editorial approach to them, though in the past week I've been drafting a replacement for
Wikipedia:"In popular culture" articles with much more specific guidance. Keep an eye on the page if you're interested.--
Father Goose (
talk)
01:50, 19 August 2008 (UTC)
No comment on whether or not its a good idea, but you should be able to make a template using {{rand}} and
WP:NAVFRAME that would allow for random display of one item with the ability to show all of them, without needing an extension or additional javascript.
Mr.Z-man15:51, 19 August 2008 (UTC)
Where would I turn to keep on top of suggestions and proposals to update the system for {whatever the currently trendy word for citations is}? I notice a huge number of modifications to the template system, but I consider these to be utterly missing the point and making the system even more of a disaster. Someone, somewhere, must be shouting "stop", but if there's an ongoing discussion about it, I can't find it.
Maury (
talk)
21:12, 19 August 2008 (UTC)
It depends what kind of changes you mean. Changes in how we use the system should go to
WP:VPR, changes in how the <ref></ref> tags actually function should go either to
bugzilla: or here if you're not sure, though if its a major change you might want to go to
WP:VPR as well.
Mr.Z-man22:32, 19 August 2008 (UTC)
Could someone have a look at
Productivity paradox. I am trying to improve citations. The link for Berglas does not work. The genrated HTML is gt cite style="font-style:normal" id="CITEREFBerglas2008" lt instead of the required <a...>. I have played with several things.
Tuntable (
talk)
04:48, 20 August 2008 (UTC)
1. I want the following table to be sortable, but it seems to be a different table rather than the common class="wikitable";
Help:Sorting does not mention how to sort these kind of tables. Can someone please make the first column- "Liquid" and the last column - "surface tension sortable". Do not make the second table sortable.
2. The second table on
List of disk partitioning software seems to have a problem. The last column "HFS" does not sort on clicking the arrows. On clicking, the view just jumps to the top part of the page. (I use Firefox 2.14). Have a look.
Regarding your second question, the problem was that the cfdisk row only had 7 cells instead of 8. Normally the table sorting script pulls the text contents of each cell, calls the toLowerCase method of the resulting string, and sorts by that. But since that final cell was missing, the "pull text contents" was getting an undefined value instead of a string and the undefined value lacks a toLowerCase method.
Anomie⚔14:38, 16 August 2008 (UTC)
Regarding the first question: you can make the table sortable by adding the "sortable" class name. However, there's only a single cell in the header row. You need to add a header for each column in order for sorting to work properly. If you want don't want to eliminate the purple header that exists now, then put this cell inside another table that wraps around the inner, data table. By having nested tables, you can have one table (e.g., the "outside" one) that isn't sortable and another (e.g., the "inside" one) that is.
SharkD (
talk)
22:12, 16 August 2008 (UTC)
The software sometimes doesn't recognise rangeblocks. For example,
208.109.0.0/16 was blocked in March 2007 as an open proxy range, yet an IP in that range –
208.109.19.19 (
talk·contribs) – managed to edit after the block.
64.34.0.0/16 was also blocked in March 2007 as an open proxy range, yet
64.34.173.112 (
talk·contribs) managed to edit after the block. So to make sure there were no other malfunctioning rangeblocks of open proxies, I went through
User:Spellcast/proxies and reblocked every range that was currently blocked. If the block form didn't allow me to reblock, that means the rangeblock is working. However, if the software allowed me to reblock, that means the old rangeblock wasn't working. After going through most of the list, I found the software wasn't recognising another five rangeblocks.
[5],
[6],
[7],
[8],
[9]. Does anyone know why the software sometimes doesn't register old rangeblocks?
Spellcast (
talk)
05:16, 18 August 2008 (UTC)
Can you confirm that the ranges were in fact blocked at the time? It's possible they got unblocked and not properly logged; or alternately that there was something borked w/ the block entry in the database. Unfortunately if there was a problem with the block itself it'll be gone due to the unblock/reblock... if you notice another one, give us a shout so we can investigate it before you try reblocking it; there might be some forensics we can do first! :D --
brion (
talk)
18:47, 18 August 2008 (UTC)
Ok here's another one (I tried reblocking every range in that subpage except for ThePlanet.com IPs, 74.222.0.0/19, and 85.234.128.0/19).
85.234.157.148 (
talk·contribs) edited this year despite the March 2007 block of
85.234.128.0/19. 85.234.128.0/19 is currently unblocked as I type despite what the block log says.
Spellcast (
talk)
00:32, 19 August 2008 (UTC)
There is no block present for that range, indicating that it was either unblocked or has expired. An unblock should have left a log entry, unless there was a bug present, and an expiry shouldn't have happened, unless there was a bug present. Either would have presumably happened at least 5 months ago, so an older bug might or might not still exist. --
brion (
talk)
13:51, 19 August 2008 (UTC)
All these rangeblocks were from 2006 to 2007 and I haven't found any for 2008. This has to be a bug, so I hope the new version of MediaWiki has fixed this. I'm assuming no new action is going to be done at this stage?
Spellcast (
talk)
05:55, 21 August 2008 (UTC)
External link searching issues
While working on a new tool I became aware the MediaWiki does not apparently index all external links. For example
North Pole links to GeoHack, however searching for the URL
yields no results. Other articles such as
Arctic Ocean do appear in the results. Is this a bug, job queuing issue, or intentional? —
Dispenser03:12, 19 August 2008 (UTC)
Figured it out. The page was blanked than ClueBot reverted it which didn't cause it to regenerate the external links table. I've check a few other edit since and those following this pattern also did not have external links. —
Dispenser13:57, 20 August 2008 (UTC)
This sounds like the parser regression which broke image resizing when there was whitespace around the size, eg [[Image:Wiki.png|80px ]], which many templates ended up expanding to. This was fixed yesterday, though some bad cached pages might yet be hiding around the corners. Just purge them if you see them. --
brion (
talk)
18:36, 20 August 2008 (UTC)
Ugh, there was so much wrong with that script (like it would break the whatlinkshere, showing only redirects to the VP/T), I had to delete it, sorry. Try this instead:
addOnloadHook(function(){varwlh=document.getElementById('t-whatlinkshere');if(wlh)addPortletLink('p-tb',wgServer+wgScript+'?title=Special:Whatlinkshere/'+wgPageName+'&hidetrans=1&hidelinks=1&limit=500','What redirects here','t-wrh','Show redirects to this page','',wlh);});
Thanks - works like a charm! Just one more small request, though: could you change the default number of results per page from 50 to 500?
It Is Me Here (
talk)
10:48, 20 August 2008 (UTC)
Hi
I contribute mostly to the French wikipedia, but felt I may receive more complete answers by asking here. Is the code to generate nice colored diff à la Wikipedia free code, and is it available somewhere e.g. as a php function? I've tried looking it up on the internet, but all I could find are some unperfect functions that would preform diffs line by line instead of work by word. Of course I could hack the function inserting return lines between every word, but it is actually slightly tricky to do, and no need to do what other people have already done.
Lerichard (
talk)
00:34, 20 August 2008 (UTC)
The default MediaWiki diff engine is
this, though Wikimedia uses a
C++ based diff engine for performance reasons. You may need some code from the various CSS files (look in the page source) for the colors.
Mr.Z-man01:04, 20 August 2008 (UTC)
thank you all for your answers, I'll take a look at it. It seems I should use a C/C++ program if I need better performance.
Lerichard (
talk)
21:39, 20 August 2008 (UTC)
Vertical alignment questions
I've been comparing the "sortable" tables in Firefox and IE and noticed that there is a difference in the vertical alignment of the clickable icon in each browser. (See
here.) In Firefox the icon is aligned to the middle of the font's lowercase letters; in IE it's aligned to the middle of the text's container. I was wondering, why is this the case, and how do I get it to behave in IE as it does in Firefox? I've also been
experimenting with how divs are rendered with varying font sizes and heights as well as presence or absence of content. I haven't reached any firm conclusions yet.
SharkD (
talk)
03:13, 20 August 2008 (UTC)
I found a setting for the vertical-align attribute, called text-middle, that wasn't documented in the reference I was looking at. It does the same trick in IE.
SharkD (
talk)
15:02, 20 August 2008 (UTC)
Too many subheadings in some genres, especially science.
I have a good education and can deal with the rather vast number of art , architecture and advertising subheadings. But I have been stymied with many science searchs because I need to know the information I am trying to find before I can find it.
I suggest some manner offering an overview. Perhaps similar to the way history is presented: overview, then an orderly motion through the subjects parts it.
Stating the meaning of mathematical expressions in English would be a help also.
I feel sure this will strike persons who got there phds in theoretical physics as a bit stupid. But there are those of use without much of a professional level mathematics background and it would be so nice to understand the thinking going on.
If there is a particular article that illustrates this issue, that would go a long ways towards us being able to address the problem.
EVula//
talk //
☯ //18:38, 20 August 2008 (UTC)
Template subst without showing all the code
Hi. Anyone knows how to use "subst:" in a template with "switch" but without showing all the code? For example:
Template:XYZ with the following content:
Example
{{#switch {{{1}}}
| 1=Text 1
| 2=Text 2
| Text 3
}}
If I use "subst:" in this template: {{subst:XYZ|2}} I get this:
Example
{{#switch 2
| 1=Text 1
| 2=Text 2
| Text 3
}}
Example Text
{{<includeonly>subst:</includeonly>#switch 2
|1=1
|2=2
|3}}
The includeonly tags prevent the substitution from being applied on the template page itself, but then cause the switch parser function to be substituted when the template is substituted. --
CBD11:20, 20 August 2008 (UTC)
(ec) If you only ever plan to substitute the template, you can put <includeonly>subst:</includeonly> into the switch statement:
Example
{{<includeonly>subst:</includeonly>#switch {{{1}}}
| 1=Text 1
| 2=Text 2
| Text 3
}}
Note, however, that this will cause corrupt output when the template is transcluded (this is the basis of the substitution check hack). If you want the template to work when transcluded as well as substituted, you'll have to use a parameter:
Example
{{{{{subst|}}}#switch {{{1}}}
| 1=Text 1
| 2=Text 2
| Text 3
}}
Here, calling {{subst:XYZ|2|subst=subst}} will give you the clean output you want. I expect this is documented somewhere, but I'm not sure where to look.
Happy‑
melon11:25, 20 August 2008 (UTC)
Thank you very much. I didn't know it was so simple. I thought I had to use a mix of noinclude, includeonly and #if. I saw documentation previously on meta but I didn't figure out.
Mosca (
talk)
15:52, 20 August 2008 (UTC)
Spanair fligh redirect incorrect with other titles
Other titles of plane crashes are AA Flight 11 Pan Am Flight 103 UA Flight 93, the title of the article has been redirected to Spanair flight AK 5022 which is wrong, it should simply be either Spani air flight 5022 or ebetter still Spanair flight 22.Pleae fix and the histories that go with the pages thanks.--
Somali123 (
talk)
15:42, 20 August 2008 (UTC)
Is there any possibility that the "Search" field on the home page can be increased in size? It's pretty hard to hit accurately using an iPhone browser. Thanks!
I was told to make a bug report whenever I receive a "Database locked" message and there is not preview pane for the article's wiki markup.
Here is a screenshot. It seems I am only receiving this particular type of error page these days.
SharkD (
talk)
00:42, 21 August 2008 (UTC)