Wikipedia:Village pump (technical)/Archive 6

From Wikipedia, the free encyclopedia
Jump to: navigation, search

#ifeq and #switch

I've seen the outputs of these templates, but what exactly is the difference? They do the same thing except #switch does that and more. —Coastergeekperson04's talk@11.21.2007 04:57

#switch could completely replace #ifeq (although #ifeq is shorter), and nested #ifeqs could replace #switch. When you have a true/false choice, #ifeq may be more clear and conceivably could be slightly faster (probably not though, or at least not enough to matter). #switch is certainly more clear if you have multiple options. Anomie 05:15, 21 November 2007 (UTC)
They aren't quite identical; imagine trying to compare two strings which contain the = character. #ifeq can do it, #switch can't. --ais523 12:05, 21 November 2007 (UTC)
Good point, I hadn't thought of that. Anomie 12:18, 21 November 2007 (UTC)
Neither had I, until the problem actually came up a couple of days ago, and in the worst possible namespace for it as well. --ais523 12:28, 21 November 2007 (UTC)

Image:Prince_symbol.svg

This image is gone without a trace - there is no deletion logs, no nothing - what's going on? —Random832 15:56, 21 November 2007 (UTC)

It was hosted at the Commons. See here and here. Lupo (talk) 16:04, 21 November 2007 (UTC)

Templates, ParserFunctions and spacing

I've been working on a few templates recently and have had issues using ParserFunctions which either insert extra spacing or break the formatting of a table. (The templates are Template:Infobox Canada electoral district, which has extra spacing after the image, and the test template User:Mindmatrix/Parliament Infobox). For example, in the latter case, I have the following code:

|-
! colspan="3" | Sessions
|  
{{ #if: {{{session-1-start| }}}
  |
  {{!-}}
  {{!!}}      First
  {{!!}} {{{session-1-start}}} - {{{session-1-end}}}
}}
{{ #if: {{{session-2-start| }}}
  |
  {{!-}}
  {{!}}      Second
  {{!!}} {{{session-2-start}}} - {{{session-2-end}}}
}}
{{ #if: {{{session-2-start| }}}
...

which produces extra whitespace at the end of the template. If I collapse it like so:

|-
! colspan="3" | Sessions
|  
{{ #if: {{{session-1-start| }}}
  |
  {{!-}}
  {{!!}}      First
  {{!!}} {{{session-1-start}}} - {{{session-1-end}}}
}} {{ #if: {{{session-2-start| }}}
  |
  {{!-}}
  {{!}}      Second
  {{!!}} {{{session-2-start}}} - {{{session-2-end}}}
}} {{ #if: {{{session-2-start| }}}
...

the spacing issue disappears, but the table breaks (adding each #if as an extra column instead of an extra row). Am I missing the obvious, or is there something quirky going on? Mindmatrix 17:29, 21 November 2007 (UTC)

First, you shouldn't place the {{!!}} (||) cell seperators on seperate lines, as they are intended to be used on the same line. Second, parser function are sensitive to line breaks, thus the function should be on one line, or comment out the line breaks using <!-- and -->.  – EdokterTalk 18:28, 21 November 2007 (UTC)
Yeah, I've noticed how sensitive parser functions can be to line breaks. I've used the nowiki hack mentioned below, though the comment markup you suggest should work just as well. Thanks. Mindmatrix 20:27, 21 November 2007 (UTC)
That's not the point. Use:
|-
! colspan="3" | Sessions
|  
{{ #if: {{{session-1-start| }}}|
 {{!-}}
 {{!!}}      First
 {{!!}} {{{session-1-start}}} - {{{session-1-end}}}
}}<nowiki/>
{{ #if: {{{session-2-start| }}}|
 {{!-}}
 {{!}}      Second
 {{!!}} {{{session-2-start}}} - {{{session-2-end}}}
}}<nowiki/>
...
{{ #if: {{{session-10-start| }}}|
 {{!-}}
 {{!}}      Tenth
 {{!!}} {{{session-10-start}}} - {{{session-10-end}}}
}}<nowiki/>
|}<noinclude>
(Worked for me.) — Komusou talk @ 18:37, 21 November 2007 (UTC)
Thanks; the <nowiki /> tag did the trick. Mindmatrix 20:27, 21 November 2007 (UTC)

Disambiguation in sortable tables?

Hi, how do you go about disambiguating names in sortable tables, such as the one at http://en.wikipedia.org/wiki/List_of_Leeds_United_A.F.C._players (Paul Robinson -> Paul Robinson (goalkeeper) ) I've tried a few things in the sand-box, but nothing seems to work.

Thanks 86.130.126.118 (talk)

The documentation for Template:Sortname is fairly clear about that. Just add the dab page as a third parameter. Andrwsc (talk) 21:51, 21 November 2007 (UTC)
Thanks (for sorting the dab as well), must have misread something looking through the Template. 86.130.126.118 (talk)

A different way of "archiving"?

I was about to archive my talk page for the first time, and rather than copy to a separate page, I created a historical page link as my "archive link". See User_Talk:ATren. This avoids copying large blocks of text (which I assume would save server space), keeps the history in one place, and also identifies the precise revision at which the archival took place. Does anyone else do this? It seems like a much more efficient and elegant way to archive than the typical cut/paste that most do. ATren (talk) 22:11, 19 November 2007 (UTC)

The process for doing this (either automated or manual) would be:

(1) get the historical link to the current revision. (2) edit the page, paste the historical link as the "archive" link (3) remove sections to be archived (4) save changes - if edit conflict, start over from (1).

ATren (talk) 22:17, 19 November 2007 (UTC)

I've noticed a few user talk pages archived like that in passing, though I can't think of an example offhand. A major disadvantage is that your archives become impossible to search (other than manually, by pulling up each page and searching it within your browser); this becomes tiresome once you have a half dozen or so archives, and completely impractical once you get up to twentyish. Server space isn't really an issue: unless things have changed a great deal since the last time I looked at the relevant part of Mediawiki's source (about six months ago), or I misread it completely, the full text of each revision is saved individually - each edit of a page takes up very nearly the same amount of storage as pasting it onto another page. There's some discussion of this at Help:Archiving a talk page#Permanent link archives method. —Cryptic 22:55, 19 November 2007 (UTC)
Ahhh, there it is. Thanks. I knew (a) someone must thought of it and (b) there were probably reasons why it's not done universally. :-) Thanks for the tips. ATren (talk) 23:02, 19 November 2007 (UTC)
Also, there's really no reason to worry about performance. Unless you're archiving each and every thread, it'll be fine. EVula // talk // // 23:06, 19 November 2007 (UTC)
I'm one of those who use this archiving method, and I also try to keep my talk page as empty as possible. I know about "don't worry about performance" slogan, but I still find that the system where I'm fixing one little mistake and the server has to save hundreds of KB (on large forum pages) very inefficient. Mediawiki-style forums is not something the devs should be proud of ∴ AlexSm 23:26, 19 November 2007 (UTC)
Fortunately, a terrabyte of storage costs only a few hundred dollars, case included, these days. – John Broughton (♫♫) 00:26, 23 November 2007 (UTC)

HANNIBAL BIRTH PLACE

Youth (247-219) When Hannibal (in his own language: Hanba'al, "mercy of Baal") was born in 247 BCE, his birthplace Carthage (today a suburb of Tunis, the capital of Tunisia) was losing a long and important war. It had been the Mediterranean's most prosperous seaport and possessed wealthy provinces, but it had suffered severe losses from the Romans in the First Punic War (264-241). After Rome's victory, it stripped Carthage of its most important province, Sicily; and when civil war had broken out in Cartage, Rome seized Sardinia and Corsica as well. These events must have made a great impression on the young Hannibal. VVVVVVVVVVVVVVVVVVV why does this say TUNISIA & WIKIPEDIA says SPAIN?

I have undone the false edit.[1] PrimeHunter (talk) 03:39, 23 November 2007 (UTC)

Searching and redirects

There is an article for BF theory, which redirects to BF model. If I search for "BF theory", I obviously want the "BF model" article. But it seems Wikipedia's search doesn't index the titles of redirect pages, which is crazy: a search for "BF theory" gives lots of completely unrelated pages, with "BF model" only appearing a long way down the list! This needs to be sorted out. JamieVicary (talk) 21:01, 19 November 2007 (UTC)

If you enter "BF theory" into the search box and click "Go" (not "Search"), you'll go to BF model. If you put "BF Theory" into the search box (note the difference in capitalization), then the "Go" fails and the displayed search results aren't very good. That's because Wikipedia's internal search engine (free, open source) is widely acknowledged to be fairly poor. You're better off searching Wikipedia via Google, Yahoo, or MSN, using site=en.wikipedia.org in the search to limit the results. (And it would help if the phrase "BF theory" actually appeared in the BF model article.) – John Broughton (♫♫) 01:48, 23 November 2007 (UTC)
I believe the problem here is that Wikipedia search has apparently stopped searching in redirects. So a Search (not Go) on "BF theory" doesn't discover the redirect at BF theory to BF model, and the latter doesn't mention "BF theory". PrimeHunter (talk) 02:49, 23 November 2007 (UTC)
Pressing "go" for 'BF Theory' seems to take me to a blank page. Voice-of-All 03:04, 23 November 2007 (UTC)
Not sure what you mean by "blank page"; for me, the "Go" takes me to a page that says "No page with that title exists". That's because of the capital "T" for the second word; as we know, Wikipedia is persnickety about interior capitalization.
As for stopped searching in redirects, is there some indication somewhere that it ever did this? – John Broughton (♫♫) 14:38, 23 November 2007 (UTC)
I don't have evidence but I clearly remember seeing lots of redirects in searches earlier, maybe a few months back. The search results page showed that it was redirects. I don't know why it stopped. PrimeHunter (talk) 15:17, 23 November 2007 (UTC)
You're right: Wikipedia talk:Searching/Archive2#Redirect pages in search results is a May 2007 complaint that for some searches, both the Go and the Search button will lead you to a search results page which redundantly includes all redirect pages. – John Broughton (♫♫) 01:16, 24 November 2007 (UTC)
Scholarpedia has a "List redirects" search option (off by default). For example, http://www.scholarpedia.org/article/Special:Search?ns0=1&redirs=1&search=manifold+theorem&searchx=Search finds a redirect but finds nothing if "List redirects" is unchecked. (But their own search generally sucks). The most useful would maybe be that by default a redirect is listed if and only if the target is not listed. That would have worked for BF theory. PrimeHunter (talk) 01:50, 24 November 2007 (UTC)

Formatting of a list

In this edit of this article the format of a list was changed. I think the new format is ugly and uncommon, and was wondering if I should change it back? I notice that WP:LIST says that these lists should be reserved for term/definition pairs, but then uses them as a sort of substitute for minor headings (as does the linked article). SharkD (talk) 19:43, 21 November 2007 (UTC)

That's an absolute misuse of the semicolon; semicolons are intended to be for minor headings that keep the headings out of the TOC. These aren't headings, because there isn't enough text following them (say, a paragraph). There is absolutely no reason why this should be anything but a bulleted list, and the titles shouldn't be bolded either. – John Broughton (♫♫) 00:19, 23 November 2007 (UTC)
Help:Wikitext examples#Organizing your writing says semicolons are for definition lists, not headers. Anomie 15:15, 23 November 2007 (UTC)
The cited page says nothing about whether semicolons can be used to as a way of avoiding a bloated TOC, it just says that one use is for definition lists. And in fact (I'll argue) the definition list has, among other purposes, avoiding making every word that is being defined into a header (again, bloating the TOC), while giving it a header-like format.
More to the point, in the article originally mentioned, a bulleted list would work absolutely fine; there was no need to get creative. – John Broughton (♫♫) 01:09, 24 November 2007 (UTC)

Last edit to a page appears before the second-last edit?!

I noticed something funny when unprotecting Image:Formidable Sydney Boom (AWM P00444-047).jpg using WP:TW. In the image's history, it says I unprotected first, then removed the template. But my contribs list shows that the template was removed first, and the unprotection happened later, and curiously, the "(top)" edit is the earlier one. How could that be? :-) Awyong Jeffrey Mordecai Salleh 00:31, 23 November 2007 (UTC)

Your contribs list also shows unprotection first. Maybe you misread it, or maybe there was a glitch that has since fixed itself. (Contribs lists are newest at the top.) --ais523 12:49, 23 November 2007 (UTC)
I recall noticing things like this before occasionally; one incident that sticks in my mind is an entry in Special:Recentchangeslinked where the vandalism reversion was showing "before" the vandalism. Every time, the out-of-order actions were done in the same minute; I wonder if something in the code is sorting by that value instead of the full timestamp or something. Anomie 22:35, 23 November 2007 (UTC)

The Age of Enlightenment

the infobox on the page The Age of Enlightenment seems buggy. Part of the information appears outside the box as if it linked to categories. I think the {{Age of Enlightenment}} is buggy.--Anne97432 (talk) 11:15, 23 November 2007 (UTC)

Watch  : [2]

The problem is that {{Navbox}} only supports up to group20. {{Enlightenment}} currently tries to use Navbox up to group23, so it fails after group20 which is Switzerland. PrimeHunter (talk) 12:06, 23 November 2007 (UTC)
I have commented out 3 groups [3] until somebody else finds a better solution. PrimeHunter (talk) 12:27, 23 November 2007 (UTC)
It looks like you're using a huge navbox when a category would have been more appropriate. – Tim Starling (talk) 12:41, 23 November 2007 (UTC)

Semi-protected pages disappear from watchlist

Is it a known bug that articles that are semi-protected disappear from watchlists? I thought I had seen this happen before but I finally took the time to verify it's happening to me right now. I edited an article this morning to remove some vandalism. I then requested the article be semi-protected as the vandalism was frequent. After it was semi-protected, the article disappeared from my watchlist. If I view the list of articles I have watchlisted, it's still listed there. It's still listed in my contributions. --ElKevbo (talk) 20:49, 23 November 2007 (UTC)

And it's not all articles as George W. Bush still appears on my watchlist despite being semi-protected. --ElKevbo (talk) 20:50, 23 November 2007 (UTC)
If you look at its article history, the latest entry there is the adding of semi protection. Since this isn't actually an edit, it doesn't get put on the watchlist and therefore the page disappears from the watchlist. This is a bug which should hopefully be fixed eventually. Fortunately, the next time the page is edited properly, there will be an edit to appear on the watch list so you shold see the page back on it again. Tra (Talk) 21:01, 23 November 2007 (UTC)
Gotcha. Are the devs aware of this bug? Is it documented? --ElKevbo (talk) 21:09, 23 November 2007 (UTC)
There's bugzilla:5546 which I think covers it. Tra (Talk) 21:28, 23 November 2007 (UTC)
It's pretty well-known. Some admins add the pp template after protection; I assume this is a sort of workaround. Gimmetrow 21:30, 23 November 2007 (UTC)

Please Help?

Hi, I tried to download Twinkle and Friendly on my monobook, but it hasn't been working. All it says on the bottom of my computer screen is "error on page". Could anybody tell me what's wrong??-BlueAmethyst .:*:. (talk) 21:09, 23 November 2007 (UTC)

Did you refresh your browser after installing the script? Directions for doing that are here. Or if you are using firefox, after you save the script, hold down Ctrl + Shift and press R. - Rjd0060 (talk) 22:04, 23 November 2007 (UTC)
No luck.-BlueAmethyst .:*:. (talk) 03:49, 24 November 2007 (UTC)
Note that Twinkle will not work on IE6.  – EdokterTalk 07:21, 24 November 2007 (UTC)
I can't really picture what the problem is. Noting the IE6 comment above, I don't think Twinkle (or friendly) work with anything except Firefox, but I could be wrong. - Rjd0060 (talk) 15:51, 24 November 2007 (UTC)
IE6? I don't have Firefox, but I think I have IE6, so I believe my problem is solved. Thanks so much.-BlueAmethyst .:*:. (talk) 01:07, 25 November 2007 (UTC)
Here's a suggestion: Switch to firefox! It is a lot better in my opinion. - Rjd0060 (talk) 01:41, 25 November 2007 (UTC)
Sounds like a good idea! Thanks! :)-BlueAmethyst .:*:. (talk) 05:09, 25 November 2007 (UTC)

IP addresses

I recently saw a page edited by an IP address, but instead of the usual, xxx.xxx.xxx.xxx format I'm used to seeing it was done as xxx.xxx.xxx.xxxA. How is an "A" part of the IP address? Mbisanz (talk) 17:31, 24 November 2007 (UTC)

Which page? Could've been a deceptively-titled user. --MZMcBride (talk) 17:49, 24 November 2007 (UTC)
Here's the diff [[4]]. If its a registered user name (I don't know how to check), would this violate the user name policy? Mbisanz (talk) 20:52, 24 November 2007 (UTC)
It's unnecessarily confusing. I'd ask them to consider renaming at WP:CHU, a bureaucrat would likely perform the request for them. --Deskana (talk) 20:55, 24 November 2007 (UTC)
When in doubt, you can check the user's log.[5] "November 23, 2007 123.567.1049A (Talk | contribs) New user account". EVula // talk // // 20:57, 24 November 2007 (UTC)

To be fair, I think this is reasonably obvious - ip addresses as far as I know only go up to 255 in each of 4 sections so 567 is wrong? 86.130.126.118 (talk)

True, but not everyone knows that, including my A-Level Computing teacher. --Deskana (talk) 21:56, 24 November 2007 (UTC)

Hmmm.... The (said to be) IP-looking username in question is 123.567.1049A, with three numbers instead of four, with two of the three numbers having values greater than 255, and with an alphabetic suffix. This is not very IP-looking to me. The 123.567.1049A username was created on November 24, 2007 and used to upload this image, to assert that the uploader is the copyright holder, and to release the image into the public domain. No articles yet link to the image, but this edit by user 123.567.1049 attempted to link to it from the Diane Garnick page. A reduced-resolution version of the image appears here.

No User page has been created for the 123.567.1049A username. User_talk:123.567.1049A does exist, and contains a message from Mbisanz (originator of this discussion section) questioning whether the username is within policy. – Boracay Bill (talk) 00:23, 25 November 2007 (UTC)

I got curious and made User:PrimeHunter/IP-like user names. I haven't investigated them and I don't promise that my processing has correctly kept special characters in user names. I'm not planning to take any action based on the list. PrimeHunter (talk) 00:42, 25 November 2007 (UTC)
I saw numbers and periods and it had the vague shape of an anon. IP address, which is what threw me off. I'll agree that many of PrimeHunter's entries are much more confusing. But that list would probably be better addresed by an admin or crat. Mbisanz (talk) 00:48, 25 November 2007 (UTC)

Emulate query string with internal links?

Point: to emulate a link such as

http://en.wikipedia.org/w/index.php?title=User_talk:Komusou&action=edit&section=new

with a regular [[link]]. (For use in templates, custom signatures, etc.)

  1. According to Wikipedia:Internal query string links et al. it's not doable, but maybe...?
  2. Are there Bugzilla requests related to this? Such as:
    1. Asking to authorize {{Fullurl:}} or {{Querylink}} in sigs?
    2. Asking to extend wikilink syntax, say with a double-slash?
      [[User talk:Komusou//action=edit&section=new]]
      (Would be handy for lots of things, not having to fake links with urls and class=plainlinks).

Thanks. — Komusou talk @ 18:42, 21 November 2007 (UTC)

The markup you suggest constitutes a legitimate page title. Subpages use the slashes. Having to disable a possible page title just for aesthetic reasons wouldn't be worth it. Awyong Jeffrey Mordecai Salleh 05:42, 22 November 2007 (UTC)

Having an "external link" to en.wikipedia.org should be fine. Fullurl should be fine; the spirit of WP:SIG appears to allow this. No one needs to authorize anything, nor do we need to implement an extended wikilink syntax. <span class="plainlinks">[{{fullurl:User talk:Komusou|action=edit&section=new}} Link text]</span> should do the trick. GracenotesT § 01:12, 23 November 2007 (UTC)

  • About point 1 (and re: Gracenotes): both "Fullurl" and "subst:Fullurl" are disabled in sigs (I used the wrong word, that should be "enable" instead of "authorize"). If you try a raw sig using them, or an ext url to en.wikipedia.org, the "My preferences" gives error message "Invalid raw signature; check HTML tags". So it would require a software change to allow them. (I guess I could copy-paste a custom sig or template, but that would defeat the point of ~~~~.)
  • About point 2 (and re: Awyong): that was a quick dummy example to vizualize the thing. I'm sure if such thing was requested, the devs know how to tweak it to an acceptable syntax. More leads could include:
    • [[User talk:Komusou][action=edit&section=new&editintro=User:Komusou/editintro]]
    • [[User talk:Komusou[]action=edit&section=new&editintro=User:Komusou/editintro]]
    • [[User talk:Komusou[[action=edit&section=new&editintro=User:Komusou/editintro]]
    • [[User talk:Komusou}{action=edit&section=new&editintro=User:Komusou/editintro]]
    • [[User talk:Komusou{}action=edit&section=new&editintro=User:Komusou/editintro]]
    • [[User talk:Komusou><action=edit&section=new&editintro=User:Komusou/editintro]]
    • [[User talk:Komusou<>action=edit&section=new&editintro=User:Komusou/editintro]]
    • [[User talk:Komusou||action=edit&section=new&editintro=User:Komusou/editintro]]
  • Also, this is not for aesthetic reasons:
    • On the general use of ext links to en.wikipedia.org: it is also disabling What-links-here, and making template code longer than necessary. The ubiquitous "view-talk-edit"/"v-d-e" or the WP:DOC templates all need those faked ext links. An extended wikimarkup would be useful.
    • On the special use in a sig, I think it would be a great benefit. Even when all you want is to drop a new message to someone, you currently have to click someone's "talk" link, download his talk page, then click again the "+" tab. I was thinking of a sig such as:
          — Komusou (talk/msg)
      The point being, a single click on "msg" would now allow you to drop a new message (while still getting any useful info from the &editintro=, such as "I am away for one month"). This is not aesthetics but ergonomy, and would save both time and a little bandwith too (not downloading talk pages for every new message). Multiply by the number of times you want to drop a new message, and the number of users.
  • So, have there been Bugzilla requests for such things, wether still open or rejected? — Komusou talk @ 14:27, 25 November 2007 (UTC)

Can a template detect edit mode or editintro?

Is there some way for a template:

  1. To detect that it is displayed in preview mode during an edit?
  2. To detect that it is displayed via "&editintro=" above an inputbox?

I compared the values of {{Env}} and found nothing discriminant, so the basics are out. But maybe someone knows of a crafty hack?

Thanks. — Komusou talk @ 18:43, 21 November 2007 (UTC)

  1. {{#if:{{REVISIONID}}|This is the normal page|This is the preview}} should do it, the result of this is: This is the normal page
  2. This is a bit harder. If the template is designed to never be transcluded, you could check the value of {{FULLPAGENAME}} and if it doesn't match what the name of the template is, you can assume it's being displayed through editintro. Is there a specific example you're interested in? Tra (Talk) 19:37, 21 November 2007 (UTC)

(unindent for lot of bullets) Thanks! I had completely missed that. It's for a template that would output differently whether (1) transcluded on a saved user/talk page, (2) displayed during an editintro, (3) but would still output as on a regular page during a regular test/preview of the template itself. So, since not all cases have to be taken in count, a combination of the two could be enough:

{{#if:{{REVISIONID}}
 | <!--(1) presumed display on a saved user or talk page-->
 | <!--Previewed during a page edit OR as an edintro-->
 {{#ifeq:{{SUBJECTSPACE}}|User
  | <!--(2) presumed preview through editintro-->
  | <!--(3) presumed preview during edit/test of template code-->
 }}
}}

As for the specific example, it is a bit convoluted. It's for simplifying the editintro use of {{Usertalkback}}:

  • At the top of a user/talk page such as User talk:Komusou, the template has to output a 3rd line for "leave me a new message".
  • But if I make a direct "leave me new message" link with an editintro, my /editintro subpage has to use a duplicate of the template and its options, PLUS a special option to tell it not to display again the 3rd line (because it would look silly to tell you to click here to leave me a new message when you are already leaving me a new message).
  • So now, I was thinking of keeping the info in a single location: to have a single copy of the template and its options stored in the /editintro subpage (as well as any information such as "I am away for one month"), and transclude /editintro at the top of my talk page too. It is currently doable, but then the template will display the silly "leave me a new message" during actual editintros too. Not an awful problem, but I thought I could try to have the template autodetect whether it is (1) displayed at the top of user or talk page (normal display), or (2) displayed as an editintro when you write a new message (disable the 3rd line then), or (3) displayed as regular preview while testing/developing the template itself (a special case not to be forgotten).

So I think the code could be trimmed to this simpler thing:

{{#ifeq:"{{REVISIONID}}"{{SUBJECTSPACE}}"|"<!--PREVIEW-->"User"
 | <!-- (3) presumed editintro : no "leave me new message" -->
 | <!-- (1)(2) presumed non-editintro : keep "leave me new message" -->
}}

Of course it's not perfect and there are special cases that would belie it (such as when editing and previewing one's own talk page section 0), but that's not the normal use of the template and shouldn't matter often. Or else, I could detect if BASEPAGENAME is "Usertalkback" and such, if more accuracy is needed. — Komusou talk @ 15:28, 25 November 2007 (UTC)

More sitenotice headaches

Anyone else having a problem dismissing the fundraiser notice again? The meta: central notice has been adjusted a few more tiems, and now dismiss isn't doing anything for me (in multiple browsers / systems). Resorted to using personal css hacks again to make it go away. – xaosflux Talk 17:38, 25 November 2007 (UTC)

Disregard for now, was something specific with my logon. – xaosflux Talk 00:50, 26 November 2007 (UTC)

Number sorting

There is a problem with tablesorting at sl:wiki, I guess someone can help me here. Slovenian number writing standard is slightly different than English, we use decimal comma instead of decimal dot. So English number 1,234,456.789 is 1.234.456,789 in Slovenian. The problem arises with sorting as the software uses a different algoritm. Is there some trick in mediawiki to fix it for sl:wiki? Thanks. --Tone 19:33, 25 November 2007 (UTC)

Simply redefine function ts_parseFloat() the same way it's done in de:MediaWiki:Common.js, sv:MediaWiki:Common.js, ru:MediaWiki:Common.js and probably some others. Also mediazilla:8063AlexSm 22:28, 25 November 2007 (UTC)

Coor misbehavior, but only on Szz; same invocation works on Template:Coor/doc

A little mystery, here. I noticed that The geo coor link causes a server error when used on the page of the Szczecin Airport (SZZ), even when tested outside the infobox (without saving the page). I tried purging the cache, no luck. Basically, I get a spew ofseveral lines like these first 3:

Warning: DOMDocument::loadHTML() [function.DOMDocument-loadHTML]: error parsing attribute name in Entity, line: 49 in /home/magnus/public_html/geo/geohack.php on line 127

Warning: DOMDocument::loadHTML() [function.DOMDocument-loadHTML]: Attribute solidarno redefined in Entity, line: 49 in /home/magnus/public_html/geo/geohack.php on line 127

Warning: DOMDocument::loadHTML() [function.DOMDocument-loadHTML]: error parsing attribute name in Entity, line: 49 in /home/magnus/public_html/geo/geohack.php on line 127

...folowed by the right toolserver page.

The very same coor dms invocation when placed on Template:Coor/doc (I included it as an example in the "Geo tag" section) invokes the appropriate toolserver page without any problems. The coord also works in he Polish Wikipedia interwiki article. What's happening? I am seeing this behavior in Opera 9.24 and Safari 3.04, on the same machine (a Mac OS X 10.4.11). --Mareklug talk 00:06, 26 November 2007 (UTC)

I isolated the problem to the inability to parse the double quote HTML hex character encoding, %22, by the toolserver served up in the pagename parameter. Compare below the two cases of toolserver geohack.php URLs for the real page of the city of Zamość and the constructed hypothetical variant, %22Zamość - the first processes correctly, and the other one barfs, as in the case of Szz, which redirects to a page that contains "Solidarność" (with quotes, that is, character %22) in the article name:
This looks like a bug in the implementation of the the geohack.php URL-handling, since " is a legal Wikipedia character for article names. --Mareklug talk 04:07, 26 November 2007 (UTC)

SQL functions and lists (or as we call them disambiguation pages

Hi, I'm editing the article MS. Actually, it's more of a disambiguation page or a list. There is a little debate there about how we could format the list. I couldn't help but think of my good old database programming class. There we could pick and chose how we wanted the information to be displayed. I think it would be really handy to be able to make a table and sort it in various ways and have it display on wikipedia the way that the final user would like. For example, the article could sort the list by MS, mS, Ms, M.S., etc... or By category: medical, Aviation, etc..., then by alphabetical, etc...? I can't place 2 and 2 toghether on how SQL and a regular articles (wikipedia's database) technologies could be implemented together. --CyclePat (talk) 03:33, 26 November 2007 (UTC)

Sortable tables already exist in MediaWiki. I'm not sure how they'd work in your case though. Graham87 23:09, 26 November 2007 (UTC)
It looks like there is agreement on the talk/discussion page for this disambiguation page. Let's not go down the path of trying to format dab pages as sortable tables, please. – John Broughton (♫♫) 22:51, 27 November 2007 (UTC)
Okay! Thank you. --CyclePat (talk) 17:38, 14 December 2007 (UTC)

URGENT: Special:Recentchanges is down

It's only showing the CSS code file for it now... Nwwaew (Talk Page) (Contribs) (E-mail me) 14:45, 26 November 2007 (UTC)

Showing normal for me...  – EdokterTalk 14:48, 26 November 2007 (UTC)
Looks fine to me too. Try bypassing your cache when viewing it. --ais523 14:49, 26 November 2007 (UTC)
Yeah, normally I'd do that, but I forgot about it... *dopeslap* Nwwaew (Talk Page) (Contribs) (E-mail me)(public computer) 17:30, 26 November 2007 (UTC)

Deletion error?

I can't delete (copyvio, if anyone cares) the second version of Image:Image3.JPG. It says "There is no archived version of Image3.JPG with the specified attributes." —Random832 21:07, 27 November 2007 (UTC)

No luck here, either. —Cryptic 21:23, 27 November 2007 (UTC)
deleting the whole thing and restoring the others worked; though I'm curious as to what could have caused the error.—Random832 21:33, 27 November 2007 (UTC)

Brief fundraiser notice glitch

I just happened to glance at the fundraiser notice and found it somewhat odd. The bug seems to have been fixed already, or it may have been a local issue that fixed itself, but since I already went to the trouble of taking a screenshot and uploading it, I figured I might as well share it – maybe someone else finds it amusing, too... —Ilmari Karonen (talk) 19:14, 24 November 2007 (UTC)

I'd contact $Brion. --$Deskana ($talk) 19:16, 24 November 2007 (UTC)
Yeah, this happens on meta occasionally. Apparently just a short delay that occurs when things are updated. As I've only ever seen it happen on meta, however, I wouldn't be particularly concerned about it. AmiDaniel (talk) 01:44, 25 November 2007 (UTC)
I too have seen it on meta twice. I've never seen it here, just a minor glitch, no biggie. - Rjd0060 (talk) 01:46, 25 November 2007 (UTC)
The one I saw and took a screenshot of was right here on Wikipedia – even though it says "Meta" on the notice! Apparently whatever caused the other parts of the notice to go missing also made it default to Meta for the project name. —Ilmari Karonen (talk) 00:05, 26 November 2007 (UTC)
Oh? Hm .. I've not noticed the glitch on any wiki other than meta, but can see how it would be possible for it to occur here too. The reason the sitename appeared as "Meta" is because the text is actually generated on Meta (m:Special:NoticeText); under normal conditions, it's run through a separate processor that substitutes $counter, etc., as well as {{SITENAME}} with values relevant to the referring wiki. AmiDaniel (talk) 16:48, 26 November 2007 (UTC)
For the record, I've seen it glitch a couple times on en, as well as once on meta. Dragons flight (talk) 16:59, 26 November 2007 (UTC)
This would likely be caused by a machine floating in the server pool with an out-of-date copy of the code, either on disk or in memory. I've shut off one server which was half-crashed and may have been responsible; I'll do another sweep for bad caching options in the PHP config just to be sure. --brion 16:27, 27 November 2007 (UTC)
Fixed a server that had bad cache setting. --brion 21:44, 28 November 2007 (UTC)

category creation problem

I've attempted to create Category:Pretenders to the Libyan throne and Category:Thomas Henry Wyatt buildings (deleted old mis-capitalized/titled categories) for 2 days, to no avail. Can someone check into this, or try creating the categories?! Thanks... SkierRMH (talk) 00:10, 26 November 2007 (UTC)

You have to create category pages like other pages and save them. The page is not autocreated by placing pages in the category, although the pages are displayed before the category page is created. I have created Category:Thomas Henry Wyatt buildings. PrimeHunter (talk) 00:22, 26 November 2007 (UTC)
Thanks for that - the problem is that I am attempting to create the pages as categories, but for some reason it isn't working. I just tried to create the 1st 1 again (Libyan) and it continues to come back with "Wikipedia does not have a category with this exact name." I've probably created dozens of categories, & never had this oddity happen :-? It gets odder - I was able to create the talk page for the category Category talk:Pretenders to the Libyan throne but not the category itself! SkierRMH (talk) 03:13, 26 November 2007 (UTC)
I have also created Category:Pretenders to the Libyan throne now. Your message sounds like you tried to create it as a blank page. I created it with a parent category. PrimeHunter (talk) 03:25, 26 November 2007 (UTC)
Thanks again - I tried it both ways, to no avail :( Both times I got the same bizarre message! SkierRMH (talk) 03:46, 26 November 2007 (UTC)
A parent category isn't required: see, for example, Category:Mayors of Osaka, Japan, which was just created. (It should have a parent category, but that's another matter.) – John Broughton (♫♫) 22:56, 27 November 2007 (UTC)
It wasn't created with a parent but it did have content which was my point. When I tried to create Category:Pretenders to the Libyan throne as a blank page, it failed and I got a message like SkierRMH. PrimeHunter (talk) 21:42, 28 November 2007 (UTC)

Blockquote text size

Somehow the text size of blockquote segments has changed. Why is that, and where can we discuss it (here, perhaps)? Blockquote is sometimes used for rather long quotes - quotes that are meant to be carefully read (see Stalin's antisemitism for instance), and it is annoying to either have the normal text too large to be comfortable or the quoted text too small to be comfortable. /SvNH 02:42, 27 November 2007 (UTC)

The font size of blockquotes is 93.75% of that of the surrounding text. The size difference helps distinguish quoted text from encyclopedic prose. I don't believe the font size of a quote would cause anyone to read it less carefully, but I may be wrong. You can add the code #content blockquote { font-size: 106.67%; } to your monobook.css page to see the blockquotes in a normal font size. GracenotesT § 04:45, 27 November 2007 (UTC)
Hm, this is what it looks like (I use a rather well-known browser with a rather well-known operating system, text size set to "normal").

Blockquote.png

I wasn't implying it would cause anyone to read it less carefully - but the way it is displayed makes it harder to read it careful unless you change the text display size (which you shouldn't have to do to be able to read comfortably the contents of an article). IMHO the left margin indent serves the need to set the text apart from its surroundings well enough. /SvNH 14:10, 27 November 2007 (UTC) (Sorry for choosing a sample text with such depressing content, I didn't think of it until now).
I agree. The reduction in font size is unnecessary and causes unnecessary strain on the reader. If the intention is simply to differentiate the quoted text from that of the encyclopedia article, I think this is accomplished more effectively through changing the font type, perhaps to a Sans instead of a Serif, than through changing the font size. Personally, I find that, as SvNH said, simply indenting the text likely makes it clear that the text is quoted rather than a part of the prose. AmiDaniel (talk) 19:48, 27 November 2007 (UTC)

Hm, that doesn't look too pretty. You might find {{cquote}} to be a good (and aesthetically pleasing) alternative. As far as I can tell, there's no consensus regarding whether cquotes or blockquotes should be uniformly used for large quotations, but both appear to be acceptable. GracenotesT § 22:40, 28 November 2007 (UTC)

Not really weighing in one way or the other, but I will say that <blockquote> has been at the smaller font size for well over a year. It's an expected behavior for the font to be smaller and editors have made conscious choices based on that fact as to whether or not to use the blockquote tag. Changing it would be a significant change. --MZMcBride (talk) 00:50, 29 November 2007 (UTC)

Non-breaking spaces for dates

Is it possible to force dates to stay on one line? Trying to force my signature to stay on one line made me think of the idea, but after some further thought I can't think of any situation where I would want a date to wrap to the next line. Ideally this would be a date preference in my preferences. Is this doable/desirable? Am I missing something? Would it be possible to just force this for signatures? ~ PaulC/T+ 17:30, 27 November 2007 (UTC)

The regular expression used to detect signatures assumes there are spaces in the date; everything that parses dates from talk pages would have to be updated to handle a change like this. But I don't see much benefit to the work; it's not a big deal if a date is split across a line. — Carl (CBM · talk) 17:40, 27 November 2007 (UTC)
The css code white-space: nowrap; causes spaces to be non-breaking. The only problem with this is that it has to go around the date, so you'd have to sign with ~~~~</span>. This solution won't break regexen, but it's cumbersome, and not too much benefit results from it. GracenotesT § 17:55, 27 November 2007 (UTC)
Ah, that would be terribly involved. I was hoping for something simple like just adding a "}}" at the end of the last character for dates in signatures and prepending "{{nowrap|" at the beginning of all signatures (or the underlying wikicode in lieu of using the template). Thanks for the response. (Edit: yes, exactly what I was thinking... too cumbersome to work out.) ~ Paul (psantoraC/T+) 18:02, 27 November 2007 (UTC)

You could add ~~~~~ to your raw signature where the date goes and sign with ~~~.—Random832 18:55, 27 November 2007 (UTC)

~~~~~ in raw signatures doesn't work. One could imitate the timestamp using substituted magic words, however. GracenotesT § 01:59, 28 November 2007 (UTC)
That's what I do, anyway. I don't use nbsps, though. --ais523 11:42, 28 November 2007 (UTC)
I'm certain I've had that in my sig before - it was as a parameter to a (subst'ed) template, though, so that may have made a difference. —Random832 13:16, 28 November 2007 (UTC)

Link to nonexistent "Austimus" in Latin Wikipedia?

This change established a link from Autism to the Latin page for Autism, la:Autismus. But there is no Latin page; when I follow that link I get a missing-page diagnostic (in Latin). Wikipedia:Interlanguage links #Links to pages that do not exist suggests that I can remove the link by hand, but in the past, whenever I've edited interlanguage links, some bot soon comes along and reverts my edits. What's the best thing to do here? Eubulides (talk) 22:36, 27 November 2007 (UTC)

You can remove the link by hand, but you also need to remove the link from all other Wikipedias that contain the offending link. I checked up to Japanese and couldn't find a version of Wikipedia with that link - asking the editor who added the link as you have done is another good idea. I didn't notice the discussion at Talk:Autism before removing the link. Graham87 10:39, 28 November 2007 (UTC)
See User talk:Yurik/Interwiki Bot FAQ for more information about interwiki links - I believe most of the interwiki bots are based on similar principles. Graham87 10:47, 28 November 2007 (UTC)

Partial transclusion?

Is there a way to directly transclude part of a page, say, the first 20 lines?

I'm trying to bypass the need to place excerpts on separate pages.

(A scrollbox won't do, because the whole page is transcluded into the scrollbox, which introduces a size/load-time concern).

The Transhumanist    23:01, 27 November 2007 (UTC)

You could put the part of the page that you don't want included inside a <noinclude> tag. Would that work? —Remember the dot (talk) 23:27, 27 November 2007 (UTC)
It's for a list of the day project for displaying a featured list daily. It would be an uphill battle getting something like that approved for use on 450 lists. It would be better if it was controlled from the other end. Is that possible? The Transhumanist    10:55, 28 November 2007 (UTC)
Not with the present software. --ais523 11:41, 28 November 2007 (UTC)

Help With My Monobook.js

Whenever I put any script in my monobook.js file, the scripts never take effect. I tried purging the server's cache and bypassing my browser's cache but nothing works. My browser has JavaScript enabled and everything else enabled for Wikipedia. Is there something wrong with the scripts in my file or what? Somebody help. Parent5446(Murder me for my actions) 02:04, 28 November 2007 (UTC)

I notice you are trying to use Twinkle. Are you using Firefox? Note that Twinkle doesn't work with Internet Explorer. or any other browsers except for Mozilla Firefox. - Rjd0060 (talk) 05:55, 28 November 2007 (UTC)
Actually it will also work on Opera and Safari. —Remember the dot (talk) 06:25, 28 November 2007 (UTC)
And Camino. -Rjd0060 (talk) 07:05, 28 November 2007 (UTC)
Oh, really? I did not know that. - Rjd0060 (talk) 07:04, 28 November 2007 (UTC)
I use Firefox (don't worry, I already checked Twinkle's compatibility before I put it in, along with every other script). Parent5446(Murder me for my actions) 21:47, 28 November 2007 (UTC)

Can we start numbered list with number other than 1?

Look at Portal:Poland/Poland-related_Wikipedia_notice_board#Participants_list and you will see what I mean. There are two columns numbered via #, but the second one starts at 1 instead of 25. Can this be changed other than by converting # symbols into plain numbers?-- Piotr Konieczny aka Prokonsul Piotrus | talk 21:24, 28 November 2007 (UTC)

It's possible to do so in HTML (wide browser support):
  1. One
  2. Two
  1. Three
  2. Four
It's also possible to do using CSS (slim browser support):
  1. 1
  2. 2
  3. 3
  4. 4
However, there isn't a way to do so using wikimarkup. Cheers. --MZMcBride (talk) 22:01, 28 November 2007 (UTC)

New-page-patrolled edit oddity

I had Special:Newpages open from yesterday, unrefreshed, and followed a couple of the links for yellow highlighted article. In all cases I saw a "[Mark this page as patrolled]" link on the article. That included several articles that had already been deleted. In other words, I got a "[Mark this page as patrolled]" link on a page that says "This article can't be found".

I'm guessing this was intended as a feature, not a bug. If the software checked before displaying that link, and (say) you were working from a five minute old list, you might get frustrated hunting for the "Mark this page" link which wasn't there because someone else cleared it a minute ago. After all, there isn't any harm if multiple people "mark" a page as patrolled - so (I guess the logic goes) just show it so people can clear it, rather than frustrating them - the more eyes, the better.

But with a day-old list, as I was using, I wouldn't expect to find the "Mark this page" link in most cases; this now seems more like a bug than a feature. So - would it make sense to modify the software (if I understand how it works, and the logic) to check the date/time stamp for when a page was patrolled, if that does exist, and if it's more than (say) an hour old, NOT display the "Mark this page" link? – John Broughton (♫♫) 14:00, 28 November 2007 (UTC)

Newpages should not be showing deleted pages. It's query straight joins on page, so only slave lag would cause anything deleted to show. Deleted pages are not really useful to patrol given that. I'd leave a bugzilla report. Voice-of-All 17:40, 28 November 2007 (UTC)

bugzilla:12143. Sorry I munged the markup (guess we need a how-to somewhere). MER-C 05:15, 29 November 2007 (UTC)

Fixed with r27941, but isn't live yet. MER-C 12:40, 29 November 2007 (UTC)