Wikipedia:Village pump (technical)/Archive 10

Button in edit toolbar that could cause major problems

See this edit to list of French people, adding <noincludeonly> and </includeonly> to the start of a section heading. This caused MediaWiki's edit section feature to go haywire - in short, all section edit links after the <includeonly> and </includeonly> edited the section after the one they were supposed to. A working demonstration of this can be found at User:Graham87/sandbox2. This problem is invisible when reading a page, and would be very confusing to new users. Is there a way of detecting other instances of this issue in Wikipedia? IMO at least the includeonly link should be removed from MediaWiki:Edittools to reduce the likelyhood of this problem happening again. What appears to be a similar issue has been filed as bug 6563 and I have commented there. Graham87 14:36, 11 December 2007 (UTC)

I would not support it being removed from MediaWiki:Edittools - this is a known bug (as you pointed out) that primarily affects other things (like the problem of section transclusion mentioned in that bug post) - and shouldn't cause us to remove a useful feature that is certainly useful when used without the context of sections. There's a bug report filed, so let's let the excellent and capable developers handle the problem from here on in. :) Nihiltres{t.l} 15:22, 11 December 2007 (UTC)
The toolbar and edit tools box have been a source of much vandalism and testing, much of which has gone unreverted. The result is an encyclopedia continually filling with "Insert non formatted text here", links to example.com, and pretty images of golden spirals (with no easy way to find it all). The toolbar and edit tools are useful to many editors, but it's probably caused more problems than it has solved. Removing includeonly would take care of a more esoteric/confusing markup inclusion, but the tools are still there – of immeasurable use to constructive editors, and of immeasurable annoyance to those who actually go around and remove all the testing. GracenotesT § 16:40, 11 December 2007 (UTC)

Perhaps its a good idea if one of the bots watches for edits of this nature. It is really easy for a bot to fix a noinclude or includeonly that is basically empty. --TheDJ (talkcontribs) 19:22, 11 December 2007 (UTC)

I've submitted a request at Wikipedia:Bot requests. -- John Broughton (♫♫) 01:32, 13 December 2007 (UTC)

Cascading protection bug

When you try to edit or view pages that are under cascading protection (e.g. Red link), the edit tab still says edit this page, but it should actually say view source. - FISDOF9 03:05, 12 December 2007 (UTC)

This is due to performance reasons. Voice-of-All 23:02, 12 December 2007 (UTC)

Flag templates broken?

Some pages are screwed up (at least on my Firefox) such as Portal:Current events/Sports. --Howard the Duck 16:49, 12 December 2007 (UTC)

  Doing... cleanup work now. User:East718 recently added {{pp-template}} to all 893 country data templates, in such a way that an embedded newline was also added. Sigh. I am quickly trying to repair all this damage. Andrwsc (talk) 20:17, 12 December 2007 (UTC)
  Fixed. I think all the flag templates have been repaired now, so if you are still seeing some problems, either clear your cache, purge the specific template, or wait for the servers to catch up.... Andrwsc (talk) 20:58, 12 December 2007 (UTC)

Redirect to section

Hi,

When I type "termination shock" in the "search" field and click "Go", I'm brought to the URL http://en.wikipedia.org/wiki/Termination_shock#Termination_shock which surprises me; I had expected Heliosphere#Termination_shock. Why does this occur? Tempshill (talk) 17:00, 12 December 2007 (UTC)

It's because of the broken way redirects are implemented (they're not proper redirects; it just serves a copy of the page at both addresses, and going to a section header is a javascript hack on top of that) —Random832 17:33, 12 December 2007 (UTC)

Trouble with Preferences Gadgets.

General question,

Why have i just received a new Gadget on my preferences, the protection tool used by Administrators, on my monobook there's no script installed which includes the protection tool the only script i have installed is Twinkle, and nav popup screen. SKYNET X7000 (talk) 20:10, 12 December 2007 (UTC)

Gadgets allow users to use certain scripts without having to install them in their own monobook. You can even remove Popups from your monobook, as it is now available via the Gadget page. EdokterTalk 20:34, 12 December 2007 (UTC)
Do i have permission to use the protection tool Gadget, even though it's classed under as Administrator Gadgets. SKYNET X7000 (talk) 20:38, 12 December 2007 (UTC)
You can try... it probably just doesn't work and gives you a permission error. EdokterTalk 20:44, 12 December 2007 (UTC)
Nope, it doesn't work at all, nothing appears and no option menus for the tool are appearing, must be some random error which appears on the preferences menu. SKYNET X7000 (talk) 20:48, 12 December 2007 (UTC)
The script checks if you have a sysop flag. You could fool the script by modifiyng your monobook.js but even after that it's of not much use to you since it's main functionality is on the Protection page.
This whole conversation is why I suggested hiding sysop-only scripts ... ∴ AlexSm 21:01, 12 December 2007 (UTC)

English spelling/dialect code

I think that we should have some new code in the next revision of MediaWiki that modifys English spellings based on your IP address. That would mask my all-time pet peeve about Wikipedia, and prevent these stupid spelling debates. C'mon, it would only be a few kilobytes! Canada-kawaii 15:01, 3 December 2007 (UTC)

I would think it's rather more complicated than that, there are many cases where the appropriate translation is only clear from the context: from my own area of interest, what the British call a 'straight' on a race track, Americans call a 'straightaway'. But obviously 'straight' does not translate to 'straightaway' in other contexts. How to you automatically make these decisions? In many ways I think life would have been easier if Wikipedia had just said 'All articles to be written in US English' and those of us based elsewhere would just have to live with it!. I am not however proposing to open another debate! 4u1e 13:28, 4 December 2007 (UTC)
Agreed that this is not a trivial i18n and l10n challenge, but Wikipedia is not American and there are fundamental WP:CSB and WP:NPOV reasons to consider this. Guessing the dialect from the IP address is inaccurate since readers may be travelling or expatriate; the best approach is a user preference setting, such as the date/time settings. Also, there would need to be syntax to exempt words from adjusting to culture/dialect, and syntax to force the wordings.
However, we must get the Wiki software fixed for date autoformatting functionality, but that is another subject. Dl2000 (talk) 01:15, 8 December 2007 (UTC)
One possible solution (that I don't necessarily endorse) would be to create a syntax that would allow for specifying the substitution of words or spellings based on the language headers the user's browser sends. -- Avocado (talk) 00:20, 14 December 2007 (UTC)

Protecting non-existent pages

I have just committed code changes (r28385) which allow protection of nonexistent pages using the existing protection interface (rather than icky cascading protection hacks). In a few days, all sysops will be able to protect non-existent articles from creation by navigating to the article and clicking on the 'protect' tab (interface example here). When a user who cannot edit the article due to this protection, they will receive an error message like this. Please note that:

  • All protection under this new feature is for page creation only. When a page is created, its 'title protection' (as this is internally called) will be deleted from the database, and not replaced by standard restrictions.
  • It uses the existing protection infrastructure, so it will be logged as an ordinary protection would be, in the protection log. Additionally, as with normal protection, an expiry can be set.
  • As with most software changes, it will be live in the next few days — a schema change is required before it is activated on Wikimedia wikis.

Werdna talk 10:04, 11 December 2007 (UTC)

So when a page is created by those with privileges (aka sysops), it becomes just a normal page, editable by anyone. Am I right? Harryboyles 10:10, 11 December 2007 (UTC)
Correct. — Werdna talk 10:11, 11 December 2007 (UTC)
Will there be any central way to see which pages are protected from creation this way? In other words, will we get a Special:Protectedtitles or some such page? If not, the "icky cascading protection hacks" may remain procedurally superior in that it is currently easy to see which pages are protected from being created. Nihiltres{t.l} 15:37, 11 December 2007 (UTC)
Creation-protected pages don't seem to show up in Special:Protectedpages. Another special page like Nihiltres suggested or an option on Protectedpages would be nice. Mr.Z-man 02:01, 12 December 2007 (UTC)
What will happen to currently protected non-existent pages? sh¤y 17:29, 12 December 2007 (UTC)
The only non-existent pages currently protected are done so by cascade protection, which will continue to hold. Conversion should occur to this new system, I suspect, just as it occurred from {{deletedpage}}. GracenotesT § 18:20, 12 December 2007 (UTC)

Special:Protectedtitles has been introduced in r28506. — Werdna talk 06:25, 15 December 2007 (UTC)

Javascript Coding Question

How would one, if possible, go about changing the links at the top right? --EoL talk 21:51, 12 December 2007 (UTC)

I'm got an additional link (to User:EVula/admin, accessible also through the "1" access key) in User:EVula/monobook.js. EVula // talk // // 22:49, 12 December 2007 (UTC)

If you're talking about the user links, you can modify those by accessing their id (in the monobook skin). For instance, to change the text "my watchlist" to "my stalklist", use the following code:

addOnloadHook(function() {
    document.getElementById("pt-watchlist").firstChild.firstChild.nodeValue = "my stalklist";
});

The ids of the links are, in order, pt-userpage, pt-mytalk, pt-preferences, pt-watchlist, pt-mycontris, and pt-logout (again, in the monobook skin).

You can add a link using the addPortletLink function. It takes the arguments portlet, href, text, id, tooltip, accesskey, nextnode, which refer to the name of the "portlet" to add the item to (in this case, p-personal), what the item links to, the text of the item, a string that can be used to find the object on a page, the tooltip for the item (what you see when you hover of it), the keyboard shortcut of the link, and where to place the item in the portlet. To add a purge link to the top-right, for example, use this code:

if (wgNamespaceNumber > -1) {
    addOnloadHook(function() {
        addPortletLink("p-personal", wgServer + wgScript
            + "?title=" + wgPageName
            + "&action=purge",
            "purge", "pt-purge",
            "Request that the server re-render this page");
    });
}

If you just need to copy-and-paste code to do something like this, I'd be glad to write it. GracenotesT § 23:39, 14 December 2007 (UTC)

55. Numbered ToC headings meaningless

I received (and agree with) this complaint/inquiry via email:

I dislike the practice of a numbered contents list, correlated to un-numbered headings.
It's false information: it looks like information but isn't. In any article of size, I'll see the heading I want then to scroll down to (yes you can click, but scrolling forward to the section gives you a glance at article heft and interest points like pictures) looking for its number the contents list had just provided.
The whole point of numbers is that you can skip memorizing anything except the number of the new heading of interest. Know it's approaching by noting 5, 6, and 7 fly past, and knowing immediately that you've skipped past by seeing 9 or 10.

So, can I suggest that we add the numbers automatically to the beginning of subsection headings?

This would also help immensely on pages like this pump, which have dozens of subsections. I might glance at the ToC, see threads 11, 33, and 44 are interesting, and then quickly scroll between them.

They would obviously have to not be part of the actual hyperlinked-title, as they're subject to change. Maybe in a small/standard font, so that they're clearly unconnected from the actual heading. For this and usability reasons, I'm asking here at the VPT.

Thoughts? -- Quiddity (talk) 21:15, 11 December 2007 (UTC)

I'm tempted to write a user script that does this (not acutely) :p, but I agree: actually adding this to software would be a great idea. So long as the code creating the headers has access to what the header index is (especially with transclusion), this might be doable. GracenotesT § 00:36, 12 December 2007 (UTC)
It is there already, preference "Auto-number headings".--Patrick (talk) 01:42, 12 December 2007 (UTC)
Aha! Much thanks. -- Quiddity (talk) 06:07, 12 December 2007 (UTC)

Should we remove the numbers from the default ToC (and just use bullet points) to avoid this confusion? Or are there other good uses for the numbers in the ToC, that override this inconsistency? Thanks. -- Quiddity (talk) 20:27, 12 December 2007 (UTC)

I don't really know. However, to remove the numbers and show the bullets again for yourself, I guess you can add this CSS code to your monobook.css:
/* Hide the number */
.tocnumber {
display: none;
}
/* Show the bullet again */
#toc ul {
list-style-image: url(/skins-1.5/monobook/bullet.gif);
list-style-type: square;
margin: 0.3em 0 0 1.5em;
}
Hope this might help. —color probeTalkContribs 21:12, 13 December 2007 (UTC)
That does help, and after using it for a few days, I'm finding it vastly preferable. I quite strongly recommend that we make this the default. -- Quiddity (talk) 21:39, 15 December 2007 (UTC)

Page is not working properly

For some reason, Moshe Shmuel Glasner is not showing much of its content: large sections of the article are inexplicably missing. However, if one edits the page, the text is there. So too on Preview mode (of Edit), on Printable view, and on History of the page; in all these cases, all the text is present. Nonetheless, the ordinary view of the page is hiding much content.

The missing content is large blocks of text in sections one, four, and five; in References, and in Resources.

Three screenshots, with the error circled in red:

Despite several attempts by various individuals to fix this problem, and despite several claims that it is in fact fixed, this problem continues. These individuals looked at the screenshots above, and replied that the problem there pictured, does not occur on their computers. Nevertheless, the problem DOES occur on my computer and on several other computers that I tried. My computer has fully up-to-date versions of IE7 and Firefox.

Thank you! Sevendust62 (talk) 13:21, 12 December 2007 (UTC)

Have you tried bypassing your cache? It looks fine here in the latest version of Firefox. Additionally, I see nothing in the current HTML that should cause such a problem. Anomie 14:08, 12 December 2007 (UTC)
The cache possibility's been raised before. Didn't help. Thank you. Sevendust62 (talk) 20:24, 12 December 2007 (UTC)
The discussion last week is in #Problem with a certain page working. It seems possible some weird censoring on your Internet connection is blocking paragraphs containing certain words like o - r - a - l. Try comparing the end of the linked section to what you see in the edit box. PrimeHunter (talk) 14:39, 12 December 2007 (UTC)
If you cannot see a preceding paragraph by me then try editing this section. PrimeHunter (talk) 14:39, 12 December 2007 (UTC)
I see the preceding paragraph a-okay. But your possibility of internet censorship is promising. I am indeed working via a kosher ISP that blocks out supposedly unkosher websites, via some process that defies rational classification. I've only been aware of it blocking entire websites, however; so far as I am aware, it does not block parts of a given HTML page without blocking the entire thing, along with a "This page is forbidden, etc." screen. But the possibility is worth pursuing. Thank you. Sevendust62 (talk) 20:24, 12 December 2007 (UTC)
I am typing the word "oral". Let's see if this shows up or not. Sevendust62 (talk) 20:24, 12 December 2007 (UTC)
Wikipedia is a famous and popular site that ISP's would probably not want to block unless their government forces them. But Wikipedia is also infamous for vandalism, often in the form of obscene words inserted in otherwise serious articles. I haven't heard about it but I could imagine a "kosher ISP" would try to avoid this problem by blocking sentences or paragraphs with certain words (maybe only in texts identified as being related to judaism?). I think all paragraphs which are blocked in your images and descriptions contain the word o - r - a - l, except one: The first paragraph in Moshe Shmuel Glasner#Method of Study. What would tricker a block there? The word s - t - r - a - i - g - h - t? That would be odd. I don't know the hebrew looking words. Which ISP is it? PrimeHunter (talk) 23:14, 12 December 2007 (UTC)
It's an ISP called "Rimon", I think. It's Hebrew language but based in England, I think. I'm not exactly sure, because my Hebrew is still rudimentary.

Update:I have just determined that only http://en.wikipedia.org/wiki/Moshe_Shmuel_Glasner is not working; http://en.wikipedia.org/w/index.php?title=Moshe_Shmuel_Glasner works just fine. The former (problematic) one is the page arrived at via the search box (search for "Moshe Shmuel Glasner") and internal linking (i.e. [[Moshe Shmuel Glasner]]). Can anyone else confirm? So it would seem the problem is NOT due to my computer or my ISP, as the first URL does not work but the second URL works just fine.Sevendust62 (talk) 00:52, 14 December 2007 (UTC)

I have just found the Eliezer Berkovits page is likewise: /wiki/Eliezer_Berkovits is missing section four, but /w/index.php?title=Eliezer_Berkovits works just fine. Sevendust62 (talk) 00:54, 14 December 2007 (UTC)
So Eliezer Berkovits, another article about a rabbi, is missing the section with o - r - a - l in the title. I see it without problems.
A rabbi is behind Rimon Internet. From http://www.ynetnews.com/articles/0,7340,L-3446129,00.html:
"Glatt kosher internet. Rimon Internet offers dynamic system for blocking inappropriate sites at supplier level, includes real-time inspection of new sites. ... it analyzes any internet site that the user is trying to access and "decides" whether or not to permit access. This analysis takes place at real-time and at internet supplier level. ... the company offers five programs with different filter degrees and a centre where sites, unfamiliar to the system, are manually checked out in real time and decided whether or not they should be censored."
They apparently started a few months ago. I haven't found details of their censoring and maybe they are still experimenting but I definitely guess they are behind your missing paragraphs. Maybe they know that http://en.wikipedia.org/wiki/... is for static reading, while http://en.wikipedia.org/w/... is for interaction where censoring can have chaotic consequences. PrimeHunter (talk) 01:42, 14 December 2007 (UTC)
Maybe. Perhaps I need to have a talk to them about their methods...interesting thing is, the entire quote you posted from Ynet, s.v. "Glatt kosher internet" through the end of the quote, doesn't show up. Neither does the link itself (to the Ynet article). But everything else shows up fine from your post. It's like Rimon also filters out mentions of itself. Perhaps the censorer is a Modron and can't allow self-references. Thank you for all your help. This has been extremely difficult for me to diagnose, because I don't have any other internet available to me! Thanks for everyone's patience with me! Sevendust62 (talk) 02:02, 14 December 2007 (UTC)

I am using the same computer as before, except now at a public wireless hotspot. Everything works properly. It is apparently then the ISP and neither my computer nor Wikipedia. Sure a vexing problem though! Thanks for everyone's help. Sevendust62 (talk) 21:11, 15 December 2007 (UTC)

Suggestion for "What links here" page

Could there be a choice on the "What links here" page to select subpages? All subpages automatically link back to their parents, but there is no easy way that I've found to find all the subpages from the parent. The links automatically generated by subpages are not listed on the "What links here" page. It is very easy to create a subpage and then forget about it. There is no easy way that I know of to see what sub-pages have been created. -- SamuelWantman 07:10, 15 December 2007 (UTC)

The best way to find subpages is, for example: Special:Prefixindex/User:Sam/, or Special:Allpages/User:Sam/. Only links or transclusions that are put in by users in the wikitext are displayed on Special:Whatlinkshere, not automatically generated links like the ones on subpages. - MTC (talk) 07:41, 15 December 2007 (UTC)

Code

I followed a link to this code and i heard it might compromise my account how can i stop it from compromising my account?--Fang 23 (talk) 15:23, 15 December 2007 (UTC)

You'd need to add it to your own monobook.js for it to affect you. (every monobook.js page displays that warning by the way) Secretlondon (talk) 15:28, 15 December 2007 (UTC)

Any edits lost Friday?

I can't find several edits (mine and others) which I saw in the last few hours UTC of Friday December 14. Puzzlement expressed. -- SEWilco (talk) 04:47, 16 December 2007 (UTC)

Clicking on "You have new messages" takes me to Life

I've received the "You have new messages (Last change)" message. When I click on "new messages" it takes me to the article, Life. Is this some sort of joke? Is it a bug that needs to be reported? SharkD (talk) 05:17, 16 December 2007 (UTC)

Nevermind, it is a joke: User talk:Leranedo added a template to his talk page in order to trick users. SharkD (talk) 05:22, 16 December 2007 (UTC)
You could politely ask him to remove it. There was talk at one point of a policy against this kind of deception, which led to the single sentence at Wikipedia:User page#Simulated MediaWiki interfaces. Chick Bowen 05:38, 16 December 2007 (UTC)

WP: vs WIkipedia:

Someone has made WP:(shortcut) expand to Wikipedia:(shortcut), which is logical as it places the shortcut in the correct namespace instead of the main space. However, this change has created hundreds of redlinks, and I cannot find any justification or explanation of it, which I find inconsiderate. Anyone in the know, please explain here. Thanks Geometry guy 22:36, 12 December 2007 (UTC)

This should all be fixed soon. The [[WP:]] pseudospace is being phased out. Instead typing WP: will automatically expand to Wikipedia:. So as I understand it the links will be fixed in a while -its just a work in progress at the moment. WjBscribe 22:40, 12 December 2007 (UTC)
Please provide links to explanations. Work in progress like this should not be done on the fly without warning, as it screws everyone around. I am tempted to create the missing redirects, but I realise the correct thing to do is to move the old WP:(shortcut) pages to the new Wikipedia:(shortcut) pages to preserve the edit history. However, even as an admin, I don't have access to the old pages to do this! This spoilt an otherwise good day for me, and it has made a mess of someones RfA as well: I hope there will be no opposes because the answers to questions are full of redlinks. Please, technical people, think about the ramifications of even the most sensible changes. Thanks. Geometry guy 22:48, 12 December 2007 (UTC)
You'll find most old [[WP:]] shortcuts are already automatically relocated to [[Wikipedia:]] locations as of now. The rest will catch up soon - probably within moments. The devs have actually done this remarkably quickly. People have long complained about using [[WP:]] to create shortcuts and the issues with these cross-namespace redirects. Looks like everything has been well thought out- try and extend the devs a little AGF :-) ... WjBscribe 22:52, 12 December 2007 (UTC)
I never doubted for a moment the good faith of anyone, but just a little bit of advertising and human touch would have helped. I see this is getting done now, but I first noticed the issue nearly an hour or so ago, and that's a long time in some parts of WP. Geometry guy 23:04, 12 December 2007 (UTC)
Just checking: WT: will be repaired similarly, right? / edg 22:49, 12 December 2007 (UTC)
That's my understanding, yes. WjBscribe 22:52, 12 December 2007 (UTC)
That's my observation too, but since there is still no link to an explanation of the changes being made, we are left to guess. As I said above: inconsiderate! Geometry guy 23:06, 12 December 2007 (UTC)

And yet, some WP: redirects were not moved. How do I get to them now, to move them properly? Gimmetrow 23:00, 12 December 2007 (UTC) Will this also eventually happen to CAT:, T: and P: redirects? Gimmetrow 23:03, 12 December 2007 (UTC)

Several WP still not working. DuncanHill (talk) 23:03, 12 December 2007 (UTC)
I suggest to wait a bit, and see if there is still a problem. Geometry guy 23:05, 12 December 2007 (UTC)
It's bizarre - if I click on the red shortcuts on my userpage, I get taken to an edit screen with text already in it. DuncanHill (talk) 23:08, 12 December 2007 (UTC)
That's a caching issue, probably. None of the WP: links on your page look like redlinks to me. I discovered this update with WP:GAN. Someone else had created a page which I've made a redirect. So there's a page in the way now, and if it's deleted, this fairly common shortcut won't work until this all is fixed. Gimmetrow 23:11, 12 December 2007 (UTC)
I just did a null-edit on my userpage and they turned blue. DuncanHill (talk) 23:16, 12 December 2007 (UTC)
Special:Statistics says the job queue length is currently 870,000+; think of all of those as being pages needing an automated null-edit to fix a caching issue. For comparison, de:Spezial:Statistik says the job queue length over there is 148. —CComMack (tc) 12:40, 13 December 2007 (UTC)

What happens in situations in which there were already redirects at the "new" location (e.g., WP:OR and Wikipedia:OR, which were formally two different pages)? Which page history is maintained? I'm asking this because I recall there are a few pages where the Wikipedia: page was not a redirect... --- RockMFR 23:11, 12 December 2007 (UTC)

Ahh, I've answered my own question. Wikipedia:FAQ looks alright. --- RockMFR 23:17, 12 December 2007 (UTC)
The change appears to have been done well. However, I came to VPT expecting to find a thread on what was going on. I did not expect to have to start one! I hope that there is something to be learnt from this. Geometry guy 23:23, 12 December 2007 (UTC)

Something that sucks about this is it's impossible to see what links to the WP shortcut. There are probably tons of links to WP:RFA, but not nearly as many to Wikipedia:RFA.[1] EVula // talk // // 23:22, 12 December 2007 (UTC)

As far as I can see, the "what links here" will eventually list them all, but those pages must be purged first. --- RockMFR 23:27, 12 December 2007 (UTC)

I find it ironic that WP:DEAD no longer works. (should go to Wikipedia:Dead-end pages). -- Kendrick7talk 23:22, 12 December 2007 (UTC)

It works fine for me... Ryan Postlethwaite 23:23, 12 December 2007 (UTC)
Yep: by the time I wrote that, the dead dead link to the dead ends project was alive alive again.... -- Kendrick7talk 23:25, 12 December 2007 (UTC)

Still broken for logged-out users. When logged in, this problem goes away entirely for me. When logged out, it is firmly in view. For example, logged in all WP: links at Wikipedia:Requests for arbitration/Matthew Hoffman/Proposed decision are blue. Logged out, they are uniformly red (even though clicking does give me an edit box containing a redirect). Splash - tk 23:26, 12 December 2007 (UTC)

Its a cache issue. They were broken when I signed out, but fixed once I purged the page cache. Should be a temporary problem. WjBscribe 23:29, 12 December 2007 (UTC)
They're all cache misses. Please add ?action=purge to all URLs in which you see broken WP: space links. (Or to be more precise, add &action=purge if the URL has index.php? within it; otherwise use ?action=purge, as I said before.) Titoxd(?!? - cool stuff) 23:33, 12 December 2007 (UTC)

What's up with these --DUP pages? Category:Redirects from shortcut has a ton of them. Am I correct in believing these are where the histories have gone? --- RockMFR 23:37, 12 December 2007 (UTC)

I'll continue answering my own questions :) We need to go through the --DUP pages and make sure they are all just redirects before getting rid of them all, right? --- RockMFR 23:40, 12 December 2007 (UTC)
Right. --uǝʌǝsʎʇɹnoɟʇs(st47) 23:58, 12 December 2007 (UTC)

Doesn't this block the creation of articles starting with "WP:"? Earlier, if there were something that in real life had a name starting with "WP:", and Wikipedia wanted to make an article about it, we could. Now, these pages will just go to names starting with "Wikipedia:". Nimman (talk) 23:53, 12 December 2007 (UTC)

Here's what happened: JeLuF, a dev, made a change to the software that essentially made WP:* a redirect to Wikipedia:*. He then ran a maintenance script to check for issues in the database WP:something (recorded as 'WP:something' in namespace 0) and make it into 'something' in the wikipedia namespace. It's like a move, only without the logging. History is preserved. If this script finds any duplicates, it appends --DUP to one, and we get to go through them and make sure that the right article got saved. There shouldn't be too many of them. --uǝʌǝsʎʇɹnoɟʇs(st47) 23:57, 12 December 2007 (UTC)
When I say "not too many", I mean 934. List at User:ST47/--DUP. --uǝʌǝsʎʇɹnoɟʇs(st47) 00:06, 13 December 2007 (UTC)
Yeah, but imagine if someone writes a book called for example "WP: A history of my life". Previously, Wikipedia could write an article about this book, but now we can't, because this will instead go to a page called "Wikipedia: A history of my life". Nimman (talk) 00:21, 13 December 2007 (UTC)
Yes, and we'd have the same problem if someone wrote a book titled "Wikipedia: An Internet Revolution". All namespaces cause this problem, see Wikipedia:Naming conventions (technical restrictions)#Colon. —Remember the dot (talk) 01:50, 13 December 2007 (UTC)
Help!: A Day in the Life is a case where it was averted because the title made more sense with an exclamation point, but at one point, it was in the help namespace, and if the exclamation point weren't in the title, it'd have been a much bigger deal. Ral315 (talk) 03:16, 13 December 2007 (UTC)
We still have a redirect in the helpspace which is annoying though from a consistency standpoint. Mbisanz (talk) 03:18, 13 December 2007 (UTC)
Sure, but that problem wasn't there before. This solution was done to eliminate a namespace conflict (the main-to-Wikipedia redirects), but instead creates another one (articles can't start with WP:). Just saying that maybe this should have been thought of, and the new problem weighed against the potential benefits. Nimman (talk) 12:24, 13 December 2007 (UTC)
I guess we just have to do what we did for m0n0wall etc. The actual article will live at Wikipedia:I'm an idiot to name my book WP: (or M0n0wall) but it will appear to most users that the article is WP:I'm an idiot to name my book WP: (or m0n0wall). BTW, it appears that [[Project:]] was already the same as [[WP:]] Nil Einne (talk) 12:51, 13 December 2007 (UTC)

Template:WP:

I'm pretty sure everything at Special:Prefixindex/Template:WP: is going to need some help. All transclusions that don't specify "Template:" are broken. --- RockMFR 23:57, 12 December 2007 (UTC)

Yep. --uǝʌǝsʎʇɹnoɟʇs(st47) 00:06, 13 December 2007 (UTC)
Cleaned up {{WP:1.0}}, will do more. Gimmetrow 00:47, 13 December 2007 (UTC)
Perhaps when redirects are present, templates should link to the main page, instead of adding "template:" to link to the redirect. I doubt many people will be using the redirects if they have to add "template:" and make the link that much longer. Ral315 (talk) 03:16, 13 December 2007 (UTC)
Is there a way to make WP: not expand when used with template syntax? (not that such a question will do us much good at this point.) -- Ned Scott 03:30, 13 December 2007 (UTC)
Well, the problem is that I think eventually, the use of "WP:" will overtake the use of "Wikipedia:" within the Wikipedia namespace; at that point, people would try to transclude, say, {{WP:POST}}, and get a nonexistent template instead. I think they do that to some extent now... Ral315 (talk) 04:12, 13 December 2007 (UTC)
I can see that, but it would have made less work if it only applied to links. Actually, I think the best thing would have been to have made a WP: namespace. We think of the entire "WP:WHATEVER" as a shortcut, as an element, and this will likely cause a lot of confusion for those who try to make new shortcuts, not knowing that we'll do so by making a Wikipedia: redirect. -- Ned Scott 04:24, 13 December 2007 (UTC)
[Edit conflict] Another namespace would be excessive for this particular issue, in my opinion. What was created was a pseudo-namespace, which allows for the WP: redirects to be moved out of the (Main) namespace where they don't belong and into the Wikipedia namespace. The real issue is that users have made poor choices in giving templates names like WP:whatever instead of WP_whatever. As for the Talk:WP:whatever, a script can probably be run to move all of those into Wikipedia_talk where they belong. This change can't be reverted for just templates with the current software. Cheers. --MZMcBride (talk) 04:35, 13 December 2007 (UTC)
I was talking more about in general rather than the issue just facing the template mainspace. Another namespace might have been excessive, but on the other hand not really, since it all but already existed in how we treated WP:'s. I just think it would have been better to keep them as their own thing, for organization and simply how we think about them. Not that it really matters that much, and yeah, not that it can be changed at this point. It still would have been nice to have gotten some input on the matter before hand. -- Ned Scott 04:52, 13 December 2007 (UTC)


Also, don't forget about Special:Prefixindex/Talk:WP: - these talk pages probably should be reunited with their redirects. --B (talk) 04:20, 13 December 2007 (UTC)

Fixing... any chance this feature will be reverted for template transclusion? –Pomte 04:33, 13 December 2007 (UTC)
Goodness, I hope not. Surely, we could just get a bot to fix the templates. Heck, I've got a javascript to do it that I used for userbox migrations a long time ago (before I knew the rules on bots). --B (talk) 05:09, 13 December 2007 (UTC)
MZMcBride answered above that the answer is no. It's now important to educate editors about the few WP:-prefixed templates that have to be substituted, preferably renaming them. –Pomte 05:11, 13 December 2007 (UTC)

Special:Prefixindex/Template:Wp: also needs work, though I haven't look at the "what links here", so they might all be orphaned already. --- RockMFR 05:27, 13 December 2007 (UTC)

So.. what happened here?

Copied from above for it's own section:

"Here's what happened: JeLuF, a dev, made a change to the software that essentially made WP:* a redirect to Wikipedia:*. He then ran a maintenance script to check for issues in the database WP:something (recorded as 'WP:something' in namespace 0) and make it into 'something' in the wikipedia namespace. It's like a move, only without the logging. History is preserved. If this script finds any duplicates, it appends --DUP to one, and we get to go through them and make sure that the right article got saved. There shouldn't be too many of them. --uǝʌǝsʎʇɹnoɟʇs(st47) 23:57, 12 December 2007 (UTC)"

"When I say "not too many", I mean 934. List at User:ST47/--DUP. --uǝʌǝsʎʇɹnoɟʇs(st47) 00:06, 13 December 2007 (UTC)"


So that's what we seem to know so far.. but am I the only one wondering why no one seemed to know about this before hand? I'm not mad at anyone or anything like that, but.. that's one hell of a change without any discussion. -- Ned Scott 05:47, 13 December 2007 (UTC)

Take a look at Wikipedia:Village pump (proposals)/Archive 4#WP: pseudo-namespace. —Remember the dot (talk) 05:53, 13 December 2007 (UTC)
>:| That was called an "On-wiki consensus"? Now I am starting to get annoyed.. -- Ned Scott 05:58, 13 December 2007 (UTC)
That link points to a reasonable consensus of people for making the change; the four opposes were not even very strong ("seems like a solution in search of a problem"). There was also a thread of 20 comments going for more than a year on Bugzilla, which is where you should go if you are actually concerned about software changes. Now there is a helpful new feature (for example, I can abbreviate Wikipedia: to WP:, even if I don't know the shortcut). What is there to be annoyed by? rspeer / ɹəədsɹ 16:58, 13 December 2007 (UTC)
The comments on Bugzilla don't mean much to me when we have to clean up the mess, and when it's disconnected from the community. The on-wiki discussion was painfully short, and involved a tiny little group of editors for a very big change. That wasn't an on-wiki consensus. It's just as likely that a little bit more discussion would have still resulted in the change, but allowed us to be better prepared for it, and maybe with a log entry to explain the sudden page move. I'm not mad about what happened, and while I think another namespace would have been better, I still see this as an improvement and am thankful for it, but yes, I am annoyed. -- Ned Scott 01:09, 14 December 2007 (UTC)
I'm just amused imagining people who login to find their userpage transcluding the entirety of Wikipedia:Userboxes without any recent edits to cause that. –Pomte 06:06, 13 December 2007 (UTC)
Fixed; please revert if it diminishes your amusement in any way. GracenotesT § 06:26, 13 December 2007 (UTC)
It is pretty annoying that the moves were not logged at all. Adding a dummy edit summary "fixing WP: namespace" could have saved us all some confusion, and the edit histories would make a bit more sense now. Kusma (talk) 06:36, 13 December 2007 (UTC)

An anti-complaint

Thank you to the devs for making this much needed change. The consensus beforehand was more than sufficient, both on Wikipedia and on Bugzilla.

After a brief transition period, everything is okay, nothing was harmed, and Wikipedia has a great new feature. rspeer / ɹəədsɹ 16:52, 13 December 2007 (UTC)

While most of the cleanup is done, not quite "everything is okay" yet. Some cleanup still remains. It was on the whole a good change, but a little advance notice would have helped provide time to identify issues, discuss the best way to fix them, and get them fixed ahead of time. Gimmetrow 06:04, 14 December 2007 (UTC)

WP: question

The recent surprise change to WP: has broken a part of my userpage. [[WP:|List of WP:… abbreviations]] was a handy shortcut to a compact list of WP: abbreviations. How can I get exactly the same list that was available using WP:? None of the variations of Special:Prefixindex/something seem to be like the WP: list. For example, Special:Prefixindex/WP: gives an enormous list — much longer than the WP: list was, as far as I can remember. - Neparis (talk) 23:32, 14 December 2007 (UTC)

"WP:" and "Wikipedia:" are now the same namespace, which is why Special:Prefixindex returns so many pages. You can use Wikipedia:List of shortcuts, WP:SHORTCUTS, or WP:WP to get to the page that WP: used to get you to. —Remember the dot (talk) 23:54, 14 December 2007 (UTC)
WP:WP is certainly a fine page with its own merits, but, assuming my memory has not failed, WP didn't look like WP:WP. For example, WP:WP has a header, a contents box, several separate sections, and a separate column of explanations for the abbreviations, none of which existed in WP: as far as I can remember. I much preferred the compact layout of WP: Is it possible to reconstruct the exact contents of WP: in another namespace? Or is there a downloaded or locally cached copy of WP: anybody could point me to or copy into my userspace? - Neparis (talk) 00:11, 15 December 2007 (UTC)
There's also Category:Redirects from shortcut. I'm really not familiar with what WP: was before. I thought it redirected to Wikipedia:List of shortcuts, but apparently not. —Remember the dot (talk) 00:30, 15 December 2007 (UTC)
It used to redirect to Wikipedia:Shortcuts, then WP: was moved to Wikipedia:--DUP in the changeover then deleted as it no longer functioned as a shortcut. Tra (Talk) 00:34, 15 December 2007 (UTC)
From the Google cache of User:Neparis, I got this:
<a href="/wiki/WP:" title="WP:"><b style="color:black;background-color:#ffff66">List of WP</b>
But I don't know if that is useful or not; doesn't mean anything to me. -- John Broughton (♫♫) 16:19, 15 December 2007 (UTC)
It was useful to me. I wish I had a copy to show you exactly what the list of abbreviations was that WP: used to generate.
Are there any developers reading this who can use their knowledge of WikiMedia to work out retroactively what WP: was generating? I'd really like to recover the lost functionality of WP:. - Neparis (talk) 01:17, 16 December 2007 (UTC)
That page still exists here: Wikipedia:List of shortcuts. That's the same page -- someone just changed the layout a bit (unrelated to the recent changes). The link at [[WP:]], of course, does not. Ral315 (talk) 10:13, 16 December 2007 (UTC)
I looked at all the previous versions of that page, but I cannot find anything equivalent or even similar to what I remember of WP:. For example, none of the versions back to 6 August 2007 (the earliest date I viewed WP: in the form I remember it) looks the same. All of them are much longer and more verbose, having that large header, contents box, lots of subsections, and separate column of explanations for the abbreviations. I'd like just the original version I remember of WP: which was a very compact page with a list of the key abbreviations and corresponding links but without any explanations. I'd recreate it in my userspace if only I'd kept a copy of WP: at the time I saw it, but I didn't, so I can't, unless somebody else has a copy in their browser disk cache or saved elsewhere which they could share with me. - Neparis (talk) 20:13, 16 December 2007 (UTC)

Image resizing function

There are a lot of fair use images that need to be resized. It would be great if Mediawiki had a function to create a new version of an image in a different resolution. On the image page there could be a link "resize image". Then I could type in the new width or height and a new version of the image would be created (with the same file name). The old version should stay. It can be deleted later if necessary. --Apoc2400 (talk) 15:01, 15 December 2007 (UTC)

Since images are already resized during display on articles, I fail to see what pupose this would serve. The only thing affected would be the size of the images stored on the server. EdokterTalk 15:17, 15 December 2007 (UTC)
Yes, that's the purpose. For Fair Use images like logos we can only have low resolution images. A lot of those images are too big, so they need to be made smaller. --Apoc2400 (talk) 16:13, 15 December 2007 (UTC)
What is too big? And what is "web resolution"? For me, anything that still fits my screen (1024 screen size, max 800px img size) is considered web resolution. Plus logos are usually done as SVG, which are scalable to the infinite, so size doesn't even matter here. EdokterTalk 16:22, 15 December 2007 (UTC)
Take Image:MSN (logo).png for example. I just changed it from 1,800×716 to 500 × 199. An other case is book covers and music album covers. Basically, the image should not be much bigger than the size actually used in the article. I often use 1.5-2 times bigger so there is room for changes. I'm not of those trying to eradicate Fair Use images. Not at all. I make some images low-res so they will not be deleted. --Apoc2400 (talk) 16:46, 15 December 2007 (UTC)
That was big... instead of resizing it, take the smaller version form the source (MS has different sizes to download), saves you the work of resizing. Also, don't take the use in an articel as a guideline... like I said: anything that fits on the screen is small enough. EdokterTalk 16:57, 15 December 2007 (UTC)
Yes, in that case there is the option of downloading a smaller version from the source. You say that anything that fits on the screen is small enough. Can you motivate that? —Preceding unsigned comment added by Apoc2400 (talkcontribs) 17:02, 15 December 2007 (UTC)
Web resolution means just that; Fair-use images do not need to be postage stamp sized. EdokterTalk 00:15, 16 December 2007 (UTC)
Unless they're images of postage stamps, of course.  :-) *Dan T.* (talk) 17:16, 16 December 2007 (UTC)

You can tag images with {{fair use reduce}}, and someone will come along shortly and do the job. Rettetast (talk) 23:54, 15 December 2007 (UTC)

Contribution History

So I was playing around with the URL in "My Contributions" and discovered that by changing the variable "limit=x" one can query more than the most recent 500 entires. What would happen if someone with a fast connection and powerful comp queryed a user like BetaCommanBot with the limit set to 400,000? I'm wondering if there should be some sort of upper limit at which point the connection is refused? Here is a modified link set to 501 to show it is possible - [2] Mbisanz (talk) 08:50, 16 December 2007 (UTC)

There is a hard limit of 5000. Dragons flight (talk) 08:54, 16 December 2007 (UTC)
Ahh, thanks for the info, I figured that was probably something I shouldn't try and test out. Mbisanz (talk) 08:55, 16 December 2007 (UTC)
There's practically nothing that the average editor can do to break the site; don't worry about performance, though we do appreciate the concern. :) (admins, on the other hand, can bring it to a crawl by deleting a major page...) EVula // talk // // 17:49, 16 December 2007 (UTC)

What's with geonotice.py?

I've recently noticed that my Watchlist loads OK in my browser (both Safari and Firefox) but then the browser indicates it's trying to load "something else". For ever. It turns out that the "something else" is http://tools.wikimedia.de/~gmaxwell/cgi-bin/geonotice.py - this page appears to hang for ever. I'd be interested to know

  1. What it is
  2. Why it's stopped
  3. When it might be started again

Tonywalton Talk 13:45, 16 December 2007 (UTC)

I've commented it out, as when toolserver is down, it's usually down for weeks, even months. AzaToth 14:47, 16 December 2007 (UTC)
I wonder if there is a way to opt it out altogether. I block it with NoScripts, mostly because it makes the watchlist load way slower. -- ReyBrujo (talk) 14:50, 16 December 2007 (UTC)
Thanks for that. Tonywalton Talk 15:41, 16 December 2007 (UTC)
This is for Wikipedia:Geonotice, which is very useful in organizing something like Wikipedia:Meetup/NYC. I hope this is corrected, and that something that was working wasn't turned off.--Pharos (talk) 04:22, 17 December 2007 (UTC)
Since I am in Argentina and it is not very likely there will be an official meeting down here, it bothers me, especially when the toolserver is slow. As I said, I just block it, but would be nice to have a chance at deactivating it. -- ReyBrujo (talk) 04:27, 17 December 2007 (UTC)
See meta:Wikimedia Argentina.--Pharos (talk) 04:35, 17 December 2007 (UTC)