RationalWiki:Technical support/Archive8

This is an archive page, last updated 22 April 2013. Please do not make edits to this page.

Archives for this talk page:

<1>, <2>, <3>, <4>, <5>, <6>, <7>, <9>, <10>, <11>, <12>, <13>, <14>, <15>, <16>, <17>, <18>, <19>, < List>, <20>, <21>, <22>, <23>, <24>, <25>, <26>, <27>

, (new)(back)

I demand McAdams website to be removed

Many of you will know me, I am the man who has been trying to remove the site that defends the official version.

His website is full of errors, in fact there is enough evidence proving a conspiracy in the Kennedy assassination. http://www.mtgriffith.com/web_documents/vsmcadams.htm

“I'm tired of it, and I'm coming out with the truth. And I think it's important that people need to know who this man(Oswald) was,” --Judyth Vary Baker

"The facts are that we do not know who killed President Kennedy, that the Warren Commission named the wrong man as the assassin and never searched for the truth of the crime." -- Howard Roffman, Presumed Guilty ( 1976 ).

"Had Oswald lived to face a trial, with competent defense it would have been risky, at best, for the authorities to try to make the claim that the physical evidence conclusively linked Oswald to the Walker shooting." --Gerald D. McKnight, Breach of Trust ( 2005 ) pg. 58.

"The individual that I know as Lee Harvey Oswald I don't think had it in him to be a person capable of committing such a crime as murdering the President of the United States. I'll always believe that. The side I saw of him was a very kind and loving man, and that's the way I like to remember him." -- Buell Wesley Frazier.^{— Unsigned, by: 92.15.139.156 / talk / contribs}

We will get right on that.Tmtoulouse (talk) 16:13, 14 November 2012 (UTC)

Thanks and I also want you to put on the website that I have tired to put on instead of Mcadams one, and where it debunks the claims in Mcadams website. From a pervious version of the page. If you do not know, it is on the page called A site which defends the "official" version, critiquing conspiracy theories. http://rationalwiki.org/w/index.php?title=John_F._Kennedy_assassination_conspiracy_theory&oldid=1087565^{— Unsigned, by: 92.15.139.156 / talk / contribs}

Just curious, what do you think about Michael T. Griffith's other writings? Such as this one?

Oh, and while you are granting demands, Trent, can I have a shrubbery? Ni.--ZooGuard (talk) 16:39, 14 November 2012 (UTC)

I'll take a nine inch pianist. MDB ^{(the MD is for Maryland, the B is for Bear)} 18:09, 14 November 2012 (UTC)

OK, I'm confused

End of Answers_Research_Journal_volume_5 - does anyone else see the text "Retrieved from "http://rationalwiki.org/w/index.php?title=Answers_Research_Journal_volume_5&oldid=1106297" " (or whatever the revision ID is today) and the categories looking just like the CSS is broken? Been constant through a few page revs - David Gerard (talk) 16:00, 19 November 2012 (UTC)

I see it. Evil fascist^{oh noez} 16:05, 19 November 2012 (UTC)

Same here in Ubuntu, Chrome as well as Firefox. Sophie^Wilder 16:25, 19 November 2012 (UTC)

All better. It was an unclosed <div> tag. Cow...Hammer^time! 16:49, 19 November 2012 (UTC)

Watchlist

Sadly, mine is still broken.

This is not trivial, I consider it a disaster - my basic method for navigating recent edits to pages I care about has now been broken for months. If wikipedia can support my thousands of watched pages, why can't we? ħuman 07:59, 28 November 2012 (UTC)

Well, Wikipedia's servers and software infrastructure cost thousands of dollars a month, so that's probably the reason it works there. Furthermore, if it's only one person affected, and that person's demands are unreasonable, then it is trivial. The frank, sensible suggestion (mysteriously not raised as of yet), is that you trim your grotesquely large watchlist. The watchlist function was created to allow users to track the articles most important to them; it is not a hack to create a version of RecentChanges that's even larger and more unwieldy than RecentChanges itself. You might as well be complaining that you can't use the search function to create an instant list of every page on the wiki by searching for "the" or "a". It doesn't matter if that's an integral part of how you navigate the wiki — that's not what it's for, and you damn well know it. Radioactive afikomen ^{Please ignore all my awful pre-2014 comments.} 08:35, 28 November 2012 (UTC)

My car is a sturdy little thing, and it usually takes me where I want to go. But just the other day, I used a dozen lengths of chain to strap my apartment building to my car, so I could move it down the block a little bit. First the chains broke, and then my car's engine overheated.

I am complaining to the manufacturer that they need to fix this. Surely, if a car can carry people and household objects easily from one place to another, it must be broken if it can't transport several tons of building. Right?--^talk 10:38, 28 November 2012 (UTC)

I'm sure that someone could come up with a Greasemonkey script to filter out the 20(?) pages that Human does not want to see in RC. Генгис^{RationalWiki GOLD member} 10:54, 28 November 2012 (UTC)

Background for those not privy:As said here, it seems as though the problem is that he has about 18,800 pages on his watchlist.--^talk 11:15, 28 November 2012 (UTC)

I'm pretty sure the frank, sensible suggestion of trimming his watchlist has been suggested before. And Trent even manually deleted his watchlist once. -- Nx / talk 12:54, 28 November 2012 (UTC)

Human asked me about this on my talk page. My last question, in an attempt to diagnose the problem, was "Just how many pages are on your watchlist?" Human didn't answer, instead complaining here that it hadn't been fixed. This is rather less than helpful.

FWIW, I've been playing silly buggers with Divabot's watchlist. At 15,000 or so entries (all of mainspace and lots of image space) I got it to give me blank pages if I asked for 3000 days' changes, though not at 30 days. This also didn't quite help me duplicate the problem, as it didn't cause the out of memory error Human's did.

The problem with allocating enough memory for a pathological watchlist is that memory is the one thing we're really short of, and it would be nice not to break the server quite so much. However, with something to help diagnose stuff, at least we can try to do things.

So, Human. How many pages are actually on your watchlist? That would be a useful start, which is why I asked it before - David Gerard (talk) 21:41, 28 November 2012 (UTC)

Well, if he can't load Special:Watchlist, I'm not sure how can he find out the number. :) Perhaps by opening Special:EditWatchlist/raw, copying the entries to a text editor and looking at the number of lines?--ZooGuard (talk) 09:23, 29 November 2012 (UTC)

And even if that doesn't work, that's also useful information - David Gerard (talk) 14:58, 29 November 2012 (UTC)

As it happens, MediaWiki 1.19.3, which I've just applied, includes a fix for a watchlist bug - worth trying again - David Gerard (talk) 08:31, 30 November 2012 (UTC)

Which one is that? I can't find anything relevant in the changelog. -- Nx / talk 13:31, 30 November 2012 (UTC)

A regex thing which could theoretically break watchlists and RC. [1] Hey, bugfixes for free - David Gerard (talk) 20:55, 30 November 2012 (UTC)

Human has now sent me his watchlist! I shall be investigating what the fucking fuck ... - David Gerard (talk) 00:34, 2 December 2012 (UTC)

This one is made of pathological edge cases. Divabot can load its watchlist and Special:EditWatchlist/raw, but trying to blank it from there is throwing up PHP fatal errors and 500ing. Special:EditWatchlist gives a PHP fatal error 500 too. And it's clearly doing horrible things to MediaWiki, as the load average hits about 15 whenever I try this. Special:Watchlist/clear redirects to Special:EditWatchlist/raw. Now attempting to clear Divabot's watchlist from the database ... - David Gerard (talk) 15:56, 2 December 2012 (UTC)

And I accidentally deleted the whole watchlist for User:Huey gunna getcha‎ too. Bah. But I deleted Divabot's watchlist, and made it a copy of Human's! ... and Divabot's doesn't break. FUCK. It'll be something weird and complicated - David Gerard (talk) 16:06, 2 December 2012 (UTC)

FUCK. No, I accidentally deleted Human's. ARSE. - David Gerard (talk) 16:07, 2 December 2012 (UTC)

Well, that's a solution, isn't it? ;D--ZooGuard (talk) 16:11, 2 December 2012 (UTC)

I really didn't mean to do that. So it's a GOOD THING we both have a full list of its previous contents. I have emailed my profuse apologies with more stuff to test. (It really was completely random that it was Human.) This is what I get for doing dangerous things completely sober - David Gerard (talk) 16:12, 2 December 2012 (UTC)

Just a point of curiosity: how many pages are in Human's watchlist?--ZooGuard (talk) 16:16, 2 December 2012 (UTC)

Over 9,000 18,000. ^postate 16:33, 2 December 2012 (UTC)

Capturebot not capturing again

For at least the last six hours (as of the time of this post), Capturebot has been failing to cap and upload pictures. I put two links in the sandbox and nothing happened with them, and there are uncaptured links on WIGO:CP and WIGO:CP talk as well. However, Capturebot reports it's online when queried from the console. Ochotonaprinceps^{not a pokémon} 07:03, 5 December 2012 (UTC)

I have no idea, but will try to have a look in a few hours - David Gerard (talk) 21:36, 6 December 2012 (UTC)

Down again, since before the downtime earlier today. Peter ^{Subsisting on honey} 08:38, 23 December 2012 (UTC)

More links for JFK assassination page

moved to Talk:John F. Kennedy assassination conspiracy theory Sophie^Wilder 17:08, 25 December 2012 (UTC)

Reporting in a rude comment and reporting Brendiggg

Moved to Talk:John F. Kennedy assassination conspiracy theory‎

HotCAT

Here's something interesting: I can still add categories with HotCAT, but for some reason or another, I can no longer remove categories. Not sure if this is unique to me or not. Reckless Noise Symphony (talk) 13:20, 30 December 2012 (UTC)

Spam user creation spree

We seem to get a hell of a lot of user accounts created, most of which don't seem to make a single edit. I assume these are mostly spambot-generated. Is there anything we can or should do about it? Wēāŝēīōīď^{Methinks it is a Weasel} 23:11, 26 December 2012 (UTC)

There's various increasingly harsh actions that can be taken as and when it becomes a problem, without preemptively going all Citizendium on account creation - David Gerard (talk) 12:09, 27 December 2012 (UTC)

Obviously the security questions which were put in place a few months ago (third letter of the logo etc.) haven't been very effective. Should they be replaced with something more robust? Wéáśéĺóíď^{Methinks it is a Weasel} 15:02, 27 December 2012 (UTC)

Bump. Sprocket J Cogswell (talk) 15:30, 7 January 2013 (UTC)

There is nothing more robust outside of draconian policies like IP blocking all of china, shutting off new user creation, or requiring some sort of application procedure. Tmtoulouse (talk) 15:51, 7 January 2013 (UTC)

It seems like whatever mojo we're using to keep spambots from actually spamming the wiki are working pretty well. Let them waste their time making useless accounts. Theory of Practice "Now we stand outcast and starving 'mid the wonders we have made." 15:55, 7 January 2013 (UTC)

Night mode? ^pathetic 15:58, 7 January 2013 (UTC)

Serving stale pages

I thought this had gone away for a while, but it seems to be back: seeing a recent change, I go directly to the page in question (often the saloon) and do not see the new content until after a page refresh. Just now it was the two most recent topics which were missing, along with their responses. Since RW is the only place I see this, I doubt it is a caching thing here in me browser. If I click on the diff in RC, I see the new stuff right away. Sprocket J Cogswell (talk) 15:30, 7 January 2013 (UTC)

It is the squid cache layer, whether or not you squid updates the cache of the page depends not just on whether it has changed but also on the response time of the database server, other users accessing the page, etc. Under high load this is more likely to happen, and is actually a good thing over all because it is what keeps the site from crashing completely like it used to. Tmtoulouse (talk) 15:49, 7 January 2013 (UTC)

Uploading copyrighted images under fair use

Not sure if this is the best place to post this, but the system for uploading copyrighted images under a claim of fair use is rather messy. I can't find fair use as an option on the licensing dropdown, which means that despite filling in the {{fair use}} template I still get an angry red message telling me that the image does not have have an image copyright notice. And the "fair use upload form" (linked from the standard form) is a clusterfuck of bad code (e.g. %7CDate = %7E%7E%7E%7E%7E%0A). Wēāŝēīōīď^{Methinks it is a Weasel} 21:11, 7 January 2013 (UTC)

+1 I've had a lot of trouble with this. Can anyone direct us to the extension that does the licenses? Blue (is useful) 23:23, 7 January 2013 (UTC)

The usual explanation for why it's not on the dropdown is here. I don't know where the extension is. Peter ^{Subsisting on honey} 23:36, 7 January 2013 (UTC)

I would add a dropdown item called "Fair use (use above template)" that would simply serve as an alternative to the "None selected". It wouldn't insert a license, but it would let the upload know, so to speak, that there is a license, but it's not in the dropdown. Blue (is useful) 23:47, 7 January 2013 (UTC)

Found it: MediaWiki:Licenses is the page you want then. Don't break anything. Peter ^{Subsisting on honey} 23:50, 7 January 2013 (UTC)

Blacklist

Do we have a link-spam blacklist? aetherometry.com redirects all incoming from our site to a bestiality porn site, so I was going to add it to MediaWiki:Spam-blacklist, but that dosen't exist yet so I wasn't sure. Hipocrite (talk) 21:01, 10 January 2013 (UTC)

We use Abuse filter for that. I'll add it. Nobody^{don't bother} 21:04, 10 January 2013 (UTC)

Didn't we have a redirector for sites that pull that one? - David Gerard (talk) 21:22, 10 January 2013 (UTC)

If we did I don't know where it is. Nobody^{don't bother} 21:28, 10 January 2013 (UTC)

"New thread" button in forum

The "new thread" button in the forum portal pages (e.g. RationalWiki:Forum/Site-related discussion or RationalWiki:Forum/General discussion is busted, and brings up a "page title was invalid" page. ₩€₳$€£ΘĪÐ^{Methinks it is a Weasel} 09:38, 15 January 2013 (UTC)

Just a bad interface design behaving in a non-obvious way - if you enter a title it works, if you leave the title blank it gets confused. There should be some way to give it a better default behaviour, or at least a better failure message - David Gerard (talk) 10:16, 15 January 2013 (UTC)

number of articles in

Is there a way to get in an article the number of articles in a known category at the time of the access the article? Same for namespace?

Jean 5 5 (talk) 00:19, 17 January 2013 (UTC)

Here you go Nobody^{don't bother} 00:25, 17 January 2013 (UTC)

Thanks Jean 5 5 (talk) 01:35, 17 January 2013 (UTC)

No problem. Nobody^{don't bother} 01:48, 17 January 2013 (UTC)

Yeah no problem from the API. If you're still looking for data put your requests here and I'll get it for you. 21:44, 29 January 2013 (UTC)

Move template

Can someone add the "moved this section" template to the handy menu at the bottom of the edit window? Theory of Practice "Now we stand outcast and starving 'mid the wonders we have made." 22:10, 20 January 2013 (UTC)

I have no idea where this is set - can't find it in Special:Allmessages. Anyone else know? - David Gerard (talk) 23:16, 20 January 2013 (UTC)

On it. Ty^talk 01:08, 21 January 2013 (UTC)

Dude! Theory of Practice "Now we stand outcast and starving 'mid the wonders we have made." 01:13, 21 January 2013 (UTC)

It's what I'm here for. Ty^talk 01:18, 21 January 2013 (UTC)

Nice one. Where was that configured, btw? - David Gerard (talk) 12:17, 21 January 2013 (UTC)

MediaWiki:Edittools Ty^JFBANBSRADA 17:55, 28 January 2013 (UTC)

Wiki hiccups - it's been the squids

The squids have being arsey, so I've done something grossly inelegant and set them to restart daily (squid1 at 00:00, squid2 at 12:00) while I try to work out what's actually going wrong with them. The server itself seems mostly to have been happy, so that's nice. Service should be slightly more consistent - David Gerard (talk) 14:36, 26 January 2013 (UTC)

Trent says it's the OOM killer coming out to play. There's nothing else on said boxes except squid3. Feck - David Gerard (talk) 17:16, 26 January 2013 (UTC)

From someone who has no idea what the squids or the OOM killer are, you're doing a great job as sysadmin. Blue (pester) 17:33, 26 January 2013 (UTC)

For what it's worth, it looks like Blue and I have occasionally been hitting the API pretty hard, assuming that's how she's getting her data. 18:03, 26 January 2013 (UTC)

The API is not going to go through the squids, the issue has to do with the amount of cached material versus available memory. I think there are a few solutions but not ones I have worked with before so will take some research before implementing. Tmtoulouse (talk) 18:17, 26 January 2013 (UTC)

Have you tried combobulating the fribbit?--"Shut up, Brx." 18:06, 26 January 2013 (UTC)

Should check the flux capacitor. Blue (pester) 18:12, 26 January 2013 (UTC)

Or the continuum transfunctioner--"Shut up, Brx." 19:18, 26 January 2013 (UTC)

FUCKING THING AAAAAAAA So it broke again. I saw someone on FB flag it as failing and both squids had fallen over. I've restarted them, but evidently I'm going to need a once-a-minute checker. Fucksake - David Gerard (talk) 11:33, 28 January 2013 (UTC)

Just apply enough soy sauce to keep them twitching. :D --ZooGuard (talk) 11:40, 28 January 2013 (UTC)

Obviously PZ has sent his squishy minions out on strike - David Gerard (talk) 11:53, 28 January 2013 (UTC)

Tech blog post on the topic - will be all "axle-maxle manglebratic geeble-geep" to most - David Gerard (talk) 11:56, 28 January 2013 (UTC)

Remember the technique of shouting at it. Threatening to deep fry and serve them with a side salad might do the trick. ^narchist 13:52, 28 January 2013 (UTC)

In my professional experience, the most effective technique is to mutter a credible threat quietly. This is why, when your computer isn't working, the IT guy comes over and glares at it and suddenly it's sweetness and light until he goes away again - David Gerard (talk) 14:44, 28 January 2013 (UTC)

The threat is more credible when the utterer is holding the Sceptre of Authoritah pointed in the general direction of the offending box. In my day, that was a #2 Phillips screwdriver, preferably Xcelite, with a stinky transparent yellow handle. Sprocket J Cogswell (talk) 14:51, 28 January 2013 (UTC)

Pink flash

Today I've been seeing a transient pink background (for about a few hundred milliseconds) in the part of the window that isn't banner or sidebar. Occurrence is infrequent, and most of the time it seems to be associated with a squid-clearing refresh <F5>. Is this a feature, a bug, or retinal detachment? Just curious, but thought it was interesting. Sprocket J Cogswell (talk) 18:41, 29 January 2013 (UTC)

Has this co-occurred with any sensation of numbness or tingling, particularly in the extremities, any loss of time, trouble speaking, or dropping of the mouth? Tmtoulouse (talk) 18:54, 29 January 2013 (UTC)

Loss of time? Most certainly, but that's easily attributable to reddit. The color is roughly #ff8080, evenly from edge to edge of the field, which is the body text area of the RW page. My short-term memory is every bit as good as it was five minutes ago. I had a chicken/avocado sandwich for lunch, which I made all by myself.

By the way, I'm using the vector skin, and an older issue of FireFox. Sprocket J Cogswell (talk) 19:52, 29 January 2013 (UTC)

I cannot think of any pink element whatsoever - David Gerard (talk) 21:04, 29 January 2013 (UTC)

It hasn't happened again since, but I haven't been on the site much. If it does happen again, I'll grab more particular details. Sprocket J Cogswell (talk) 21:41, 29 January 2013 (UTC)

Gah. It was here. The hidden box has a pink background, and was taking its own sweet time to collapse. I will now give myself five minutes' time out in the drooling idiot box. Sprocket J Cogswell (talk) 22:50, 29 January 2013 (UTC)

Ya I've seen it too. I suspect just js waiting until the whole page is loaded before hiding that DOM element. It's a Good Thing that occasionally leads to a weird user experience. Lol do you always see colors in hex? 23:24, 29 January 2013 (UTC)

Now that must have just been designed to piss people off - David Gerard (talk) 00:01, 30 January 2013 (UTC)

Nutty, I see colors with the customary human wetware, but I have an image processing program handy that lets me see what various RGB values look like. Ask an artist about visual memory sometime, how it fades, how it can be fooled, and how it can be trained. I think I can remember colors more accurately and a bit longer than I can remember the sound of a particular violin, but I could be horribly wrong about both. At any rate, the GIMP tells me the pink in question is #ffc0cb, so I was horribly wrong about that too. <sob> Sprocket J Cogswell (talk) 00:18, 30 January 2013 (UTC)

JavaScript was designed to mess with users and developers. It's hard to debug and ... anything that's 99% anonymous functions with that any mustaches is ALIENS. 00:36, 30 January 2013 (UTC)

That's like demanding perfect pitch. Colors are remarkably different looking from context to context. FF8 and FFC are pretty close. The only thing that's weird is how you remembered the saturation. I can easily see people having a hard time distinguishing a washed out rosy pink and a washed out salmon even if they were presented together. Interesting how everyone remembers these sort of transitory sensory experiences differently. 00:50, 30 January 2013 (UTC)

All I can remember is reading the actual article text in the pink box and knowing that pink will be associated with said article text burnt into my brain FOREVAAAAAHHHHHH (till tomorrow, maybe the next day) - David Gerard (talk) 01:20, 30 January 2013 (UTC)

Sarcasm tag extension

I made a sarcasm tag extension here: mw:Extension:Sarcasm

I think it could be useful. We don't need to use it in articles if that would be obtrusive, but using it in discussions could be good. I also created a CSS class for the tags so it's easy to change the sarcasm text from the wiki. ఠ_ఠ Inquisitor Sasha Ehrenstein des Sturmkrieg Sector 07:59, 30 January 2013 (UTC)

Vandal bin

Is it just me, or does the "&lt;ipblocklist-username&gt;" stuff appear in the "Find a vandal binned user" box at Special:VandalBin for everyone? Peter ^{Droid whisperer} 03:05, 2 February 2013 (UTC)

I see it too. Ty^JFBANBSRADA 03:07, 2 February 2013 (UTC)

Not only that, but when you bin somebody, the first check box says "<ipbanononly>" --PsyGremlin^Snakk! 07:14, 4 February 2013 (UTC)

It appears to be in the actual PHP code for the vandalbrake extension, calling system messages that don't exist in 1.19, and this is just the first anyone's said anything. I could be wrong of course. In the meantime, I'm trying to work out what would be a good variable to use instead - David Gerard (talk) 21:05, 4 February 2013 (UTC)

I was right - fixed the first one by fiddling with the extension code, second by creating the interface message (since the message does exist on Wikipedia). Any others? - David Gerard (talk) 21:19, 4 February 2013 (UTC)

Today's Holiday on RC

It's all smooshed over to the left instead of being centred. This has happened a few times before. Theory of Practice "...and we do love you madly." 13:59, 5 February 2013 (UTC)

Fixed. Ty^JFBANBSRADA 15:09, 5 February 2013 (UTC)

I knew I could count on you, kid. Theory of Practice "...and we do love you madly." 15:16, 5 February 2013 (UTC)

Hey, it gives me something to do. Ty^JFBANBSRADA 15:19, 5 February 2013 (UTC)

WigoBot

Would someone please direct me to the code for WigoBot. Thanks. 16:09, 5 February 2013 (UTC)

Edit filter

[2] is an interesting edit filter. Implementation here should be considered. Hipocrite (talk) 21:16, 5 February 2013 (UTC)

It's our Filter 1. Blue (is useful) 21:20, 5 February 2013 (UTC)

Ridiculously insignificant bug

The editcount extension hasn't been updated since LiquidThreads was installed, so both of the LQT namespaces show up as blank spaces. Blue (is useful) 00:41, 15 February 2013 (UTC)

large scale rollback?

Is there a way to rollback all of a user's edits in one go?--"Shut up, Brx." 01:43, 15 February 2013 (UTC)

Not unless you want to write a script, though it's fairly simple if you just open the user's contribs and click "rollback - open in new tab". Blue (pester) 01:55, 15 February 2013 (UTC)

That's what I did, but my middle mouse button is deficient. Also, if the bot is prolific enough, it's possible that some of its misdeeds won't be seen. Then the vandalism will be present on an article indefinitely. Btw, thanks for activating the edit filter--"Shut up, Brx." 01:58, 15 February 2013 (UTC)

Middle mouse button? I just right click then left click on the menu. Micro! Also, how would any edits turn invisible? Blue (is useful) 02:01, 15 February 2013 (UTC)

Ctrl + left click should do the job, at least in most Windows set-ups. Wěǎšěǐǒǐď^{Methinks it is a Weasel} 02:05, 15 February 2013 (UTC)

Wow, thanks Weaseloid! I hope I remember that--"Shut up, Brx." 02:08, 15 February 2013 (UTC)

If we fail to revert quickly enough, certain edits may become obfuscated simply because they are one among so many. And I'm not a huge fan of having to use the mouse menu to open new tabs. Before my mouse got faulty, I would click the middle mouse button over a link and it would open in a new tab for me. I also have a button on my mouse that double clicks for me, and one that goes forward in my history for browsers. I are lazies--"Shut up, Brx." 02:05, 15 February 2013 (UTC)

Get popups. Blue (pester) 02:09, 15 February 2013 (UTC)

(EC) Use this list to remove most of the RC entries that are irrelvant to the cleanup. Peter ^mqzp 02:10, 15 February 2013 (UTC)

@Blue: How do you mean?--"Shut up, Brx." 02:12, 15 February 2013 (UTC)

Preferences -> Gadgets -> Browsing gadgets. First one on the list. Blue (pester) 02:14, 15 February 2013 (UTC)

They can get rather slow in this situation though. Peter ^mqzp 02:15, 15 February 2013 (UTC)

(edit conflict) I'm learning so much!--"Shut up, Brx." 02:16, 15 February 2013 (UTC)

Okay, no, popups is actually pretty annoying. Thanks anyways, Blue--"Shut up, Brx." 02:18, 15 February 2013 (UTC)

Searches exclude recent results

A few times recently I've been trying to find a fairly recent Saloon Bar discussion & just haven't been able to track it down with the barchive search. Now I see that the search just isn't picking up search terms from the last dozen or so archives at all. Random examples:

• Dyscalculia : link, search
• Monsanto : link, search
• cesspool : link, search
• Clint Eastwood : link, search
• pwnage : link, search
• CHRIIIIIIIIISTMAAAAAAAAAAASSSSS! : link, search

The same seems to be happening with the WIGO:CP talk archives:

• Teavangelicals : link, search
• Fluke : link, search
• Easter Island : link, search

Probably the same with other discussion archives. Is our search feature only able to search cached pages or something? Wėąṣėḷőįď^{Methinks it is a Weasel} 10:23, 23 December 2012 (UTC)

I had noticed that too, but I'd assumed that our seach engine was just intrinsically faulty due to MediaWiki being horribly shitty from the back end, so I didn't bring it up. Radioactive afikomen ^{Please ignore all my awful pre-2014 comments.} 10:57, 23 December 2012 (UTC)

Wikipedia doesn't seem to suffer from this problem. ₩€₳$€£ΘĪÐ^{Methinks it is a Weasel} 11:07, 23 December 2012 (UTC)

Oh come on, that's totally unfair. WP pays almost $1 million per month and employs about 3 dozen people to keep things from going to poop. Radioactive afikomen ^{Please ignore all my awful pre-2014 comments.} 11:22, 23 December 2012 (UTC)

Yeah, I was just pointing out this seems to be a site-based problem, not a general MediaWiki problem. Weaseloid^{Methinks it is a Weasel} 12:31, 23 December 2012 (UTC)

Buggered if I know. I did just restart the daemon again. Presumably it will index when it feels like it. We use the Lucene-based search, which is the same thing Wikipedia uses. Not sure what the lag is on theirs though - David Gerard (talk) 11:51, 23 December 2012 (UTC)

It seems to affect all fairly recently created pages, not just talk pages. E.g. the Errol Denton article doesn't show up on a search for Denton. ΨΣΔξΣΓΩΙÐ^{Methinks it is a Weasel} 13:37, 28 December 2012 (UTC)

Hmm ... on a few quick searches, I can't see any results past July. Can anyone else? (Trent, how tf does this thing work?) - David Gerard (talk) 13:54, 28 December 2012 (UTC)

The Lucene indexer is supposed to run every day, I think it's in cron.daily -- Nx / talk 14:02, 28 December 2012 (UTC)

I'm making this sticky (if that function is even working) till it's fixed, as this problem is making it seriously difficult to find fairly important things like the current mod & board elections. Weaseloid^{Methinks it is a Weasel} 02:35, 31 December 2012 (UTC)

574  [main] WARN  org.wikimedia.lsearch.oai.OAIHarvester  - Error reading from url (will retry): http://rationalwiki.org/w/index.php?title=Special:OAIRepository&verb=ListRecords&metadataPrefix=mediawiki&from=2012-07-21T10:25:30Z
java.io.FileNotFoundException: http://rationalwiki.org/w/index.php?title=Special:OAIRepository&verb=ListRecords&metadataPrefix=mediawiki&from=2012-07-21T10:25:30Z
       at sun.net.www.protocol.http.HttpURLConnection.getInputStream(HttpURLConnection.java:1401)

Yeah, that probably doesn't help - David Gerard (talk) 08:33, 15 February 2013 (UTC)

I tried switching on extension OAIRepository but haven't managed it quickly (the instructions are shitful). So I'm now running ./build to do a complete reindex, which is sorta not fast. Doing ./update via OAIRepository is the Right Thing, so I'll keep fiddling intermittently with that - David Gerard (talk) 08:54, 15 February 2013 (UTC)

Well, that didn't work - crashed out with:

   26231 [main] INFO  org.wikimedia.lsearch.ranks.Links  - Opening for read file path
   java.io.IOException: no segments* file found in org.apache.lucene.store.FSDirectory@file path

Fucking thing. Searches up to about July 2012 still work. Will kick further - David Gerard (talk) 08:57, 15 February 2013 (UTC)

Thank you for working on this, David! Radioactive afikomen ^{Please ignore all my awful pre-2014 comments.} 08:59, 15 February 2013 (UTC)

JUST SUCCESSFULLY BUILT THE INDEX! Now seeing if it can be done twice. (Takes 22 minutes with our 85,000 pages.) - David Gerard (talk) 21:36, 15 February 2013 (UTC)

Oh, and Trent disabled OAIRepository because it fucked up other special pages. If we want it we need to fix it properly. Urgh. - David Gerard (talk) 21:48, 15 February 2013 (UTC)

Search updated!!!

I built it and it worked. I built it a second time and it worked. This proves it works. Pretty much. So it's in weekly crontab. It took 26 minutes to run just now (for 86,584 pages), so it'll need an eye kept on it - David Gerard (talk) 22:07, 15 February 2013 (UTC)

Nice one DG. Thanks for sorting it. Weaseloid^{Methinks it is a Weasel} 22:10, 15 February 2013 (UTC)

DNSBL

The IP address of the coffee shop I work at sometimes is blocked, presumably by the edit filter. The problem is it's a Comcast static business address. If it was ever used as a proxy, it was in 2010 according to the single dated reference to the IP w/ "proxy" Google returns - that's a little weird because it's within a few months of this shop opening up but it's also possible they wouldn't have had the same static IP almost 3 years ago. There's a reference to Wed., Sept. 12, which could have been Sept. 12 last year, but this makes little sense. I was working here back then as well - it's just a bunch of hipsters and hackers. I'm interested in knowing what attempts the edit filter is actually catching and how they compared to what Brasov is able to pull off. 16:01, 13 February 2013 (UTC)

The filter isn't on, but if you give me the IP I can check the logs to see if it got caught at some point. I can go through the filter logs to give you a innocent IP:Brasov ration in a while. Ty^{Carnival time.} 16:31, 13 February 2013 (UTC)

Of the 1310 hits on the disable ip editing filter, 28 were legitamite IP edits, 6 were other vandals, the remaining 1276 were Brasov. Ty^{Carnival time.} 16:46, 13 February 2013 (UTC)

Wow. The success rate doesn't justify investigating individual IPs. 17:05, 13 February 2013 (UTC)

Eh? Ty^{Carnival time.} 17:07, 13 February 2013 (UTC)

So wait, you're using the edit filter to stop all IP editing because of one fucktard? -- Nx / talk 18:02, 13 February 2013 (UTC)

and they won't even let me run night mode... Occasionaluse (talk) 18:05, 13 February 2013 (UTC)

(EC) That was what we did when the attacks showed no sign of stopping after a few hours and we didn't want to have to wait for David or Trent to both notice and decide what to do. Currently we're employing a DNS proxy ban courtesy of David. Blue (is useful) 18:07, 13 February 2013 (UTC)

State of the Vandal

For obvious reasons it would be silly to list plans here, but we're not asleep and a number of the more technically minded editors are working on this issue - David Gerard (talk) 23:57, 14 February 2013 (UTC)

Good to know. Are all the erros in the last few minutes your "fault" or the vandal's? Peter ^mqzp 00:00, 15 February 2013 (UTC)

Tmtoulouse (talk) 00:01, 15 February 2013 (UTC) <---- fault

TOU-LOUUUUUUUUUSSSSEEEEEEE! - David Gerard (talk) 00:02, 15 February 2013 (UTC)

Thank you, David and Trent. We do appreciate all of your work, especially in times like these.

For future reference, it'd be best if you mention "we're working on it" when you start working on it, because otherwise it looks like you're ignoring the site while the rest of us struggle with damage control. You're not Seal Team 6; "operating in silence" isn't an option when other people are relying on you. Radioactive afikomen ^{Please ignore all my awful pre-2014 comments.} 00:05, 15 February 2013 (UTC)

Times like this really makes me happy that RW has the Sysop policy it does. This would be so much worse if only 5-6 people were Sysops and weren't always around. --Revolverman (talk) 00:07, 15 February 2013 (UTC)

We can all play Bag The Vandal, I nearly got one but got a 503 then PeterL got it. Bah! - David Gerard (talk) 00:28, 15 February 2013 (UTC)

They call me the botrollback ninja. Better luck next time! Peter ^mqzp 00:35, 15 February 2013 (UTC)

Yeah, that's why I thought I should say something, even as a placeholder - David Gerard (talk) 00:27, 15 February 2013 (UTC)

This is getting to be an untenable situation. I have disabled IP editing via the edit filter once again. David, Trent, Nutty, please proceed with all deliberate haste, so we can return to moderately normal functioning. Blue (pester) 01:58, 15 February 2013 (UTC)

Recent changes just gets rendered useless--"Shut up, Brx." 02:00, 15 February 2013 (UTC)

While I appreciate the work that Trent, David and Nutty are doing, I will not hesitate to activate the IP editing filter until such time as I am notified that any of their plans have gone into effect. Blue (pester) 21:17, 15 February 2013 (UTC) (By that I mean should the vandal reappear, not indefinitely.) Blue (is useful) 21:23, 15 February 2013 (UTC)

I can't promise hermetic whatever, obviously, but there's stuff happening. I just switched off the filter not to be a swinging dick, but because I am testing stuff - David Gerard (talk) 21:26, 15 February 2013 (UTC)

May I suggest that, in the meantime, the filter be set to only block mainspace (and only when the vandal is present)—thus ensuring peace of mind but not blocking testing at the sandbox. Peter ^mqzp 21:43, 15 February 2013 (UTC)

We are well past the "sandbox" testing, and its about testing in the wild now. Please don't turn off ip editing anywhere. Tmtoulouse (talk) 21:46, 15 February 2013 (UTC)

If you say so... Peter ^mqzp 21:47, 15 February 2013 (UTC)

Is it just me...

that I am getting frequent, non-persistent 403's on RW? [[User:K61824|]][[User_talk:K61824|]] 02:42, 15 February 2013 (UTC)

No, it isn't. Perhaps we should put up an intercom message? Peter ^mqzp 02:45, 15 February 2013 (UTC)

I emailed DG. I'm getting these up the wazoo--"Shut up, Brx." 02:48, 15 February 2013 (UTC)

I'd be surprised if he were still up, actually. Peter ^mqzp 03:01, 15 February 2013 (UTC)

I live in London, so those UTC timestamps are the time for me. I went to bed at 1am because the loved one told me to switch the damn computer off.

The sporadic 403s should not quite be happening, but I think I know why - David Gerard (talk) 07:38, 15 February 2013 (UTC)

I haven't seen one in a while at least. Peter ^mqzp 08:38, 15 February 2013 (UTC)

State of the vandal

So we're doing, er, stuff. The captcha has changed - please check if you can edit logged-out, report if you can or can't (Trent can, I'm finding I can't am stupid). The vandal is actually hammering the servers quite hard and very little of it is getting through, but please don't block all IPs for the moment - David Gerard (talk) 21:59, 16 February 2013 (UTC)

Cannot edit logged out. Ty^JFBAA 00:36, 17 February 2013 (UTC)

Don't just type the picture, read the instructions - David Gerard (talk) 00:45, 17 February 2013 (UTC)

*facepalm*Ty^JFBAA 00:50, 17 February 2013 (UTC)

I got caught the same, I was going "TREEEEEEEEENT IT'S NOT WOOOORKING" and then went " ... ah." - David Gerard (talk) 01:18, 17 February 2013 (UTC)

werks --83.84.137.22 (talk) 00:57, 17 February 2013 (UTC)

Test. 121.74.95.199 (talk) 01:20, 17 February 2013 (UTC)

I tried it several ways. Me and my tapeworm, we fail as a Turing machine. Ignorance is bliss, so don't bother to enlighten me. Sprocket J Cogswell (talk) 01:30, 17 February 2013 (UTC)

I just got a picture that didn't exist. Peter ^mqzp 01:34, 17 February 2013 (UTC)

Not a one-off either. Peter ^mqzp 01:44, 17 February 2013 (UTC)

I just had to do four of them in a row before my edit was saved. Blue (pester) 06:08, 17 February 2013 (UTC)

Right now it's reCaptcha - not actually great for ordinary spam (to which we've been almost immune so far), but should work for this purpose - David Gerard (talk) 08:52, 17 February 2013 (UTC)

Formatnum

Is there a way to formate numbers in the french way (199,455,333.2 => 199 455 333,2)? Jean 5 5 (talk) 18:35, 24 February 2013 (UTC)

There isn't a way right now. Extension:NumberFormat lets you do this , we don't have it installed. I would think it would be possible to format numbers using a template and some complicated parser functions, but I don't know of such a template having been written - David Gerard (talk) 19:01, 24 February 2013 (UTC)

Well, thanks for the answer, anyway, it's only a small detail. Jean 5 5 (talk) 19:05, 24 February 2013 (UTC)

WIGO:CP Talk seems to be busted

None of the signature seem to be working. --Revolverman (talk) 23:08, 24 February 2013 (UTC)

Close your capture tags - David Gerard (talk) 00:00, 25 February 2013 (UTC)

Bertran's rights at ru-rw

(moved from user talk:David Gerard)

Hello! I've posted this message to Trent, but it seems he is to busy now. There are some pages at ru-rw, that I want to change, but I have no rights to edit them. For example, this and this. Can I ask you to give me an access for editing that pages?--Mr. B 17:20, 23 February 2013 (UTC)

I can edit those pages (i.e. I can edit in MediaWiki: space), but it looks like I don't have access to Special:Userrights in ru:rw. I'll have to fiddle with the database. I'll do this later, when the kid's asleep ... - David Gerard (talk) 17:54, 23 February 2013 (UTC)

There is no need to access to Special:Userrights in ru:rw. User rights management in ru.rw is performing at en.rw.--Mr. B 21:56, 24 February 2013 (UTC)

That means giving you moderator rights, which is the subject of stupidity at en:rw ... I'll just be a moment - David Gerard (talk) 21:59, 24 February 2013 (UTC)

No it doesn't - you've already got the Tech right, which includes the editinterface right ... so you should be able to edit those pages ... Can you edit the en:rw equivalents? Anyone else - what's going weird here? - David Gerard (talk) 22:08, 24 February 2013 (UTC)

By the way, yes! I can edit en-equivalents! But none of those ru-pages.--Mr. B 08:30, 25 February 2013 (UTC)

I suggest that removing my tech-rights and returning them may give me an access to those pages. What do you think about this?--Mr. B 17:59, 25 February 2013 (UTC)

I've just tried that. Can you get in?

Also, to solve your immediate problems: please create the new versions of the ru:rw pages you want to change, somewhere that I can copy them to the correct places, and I'll put them there.

Trent says the ru:rw configuration and setup is just completely messed up. I'll try to find time to look at it properly ... - David Gerard (talk) 20:20, 25 February 2013 (UTC)

OK, the 'tech' group had no powers at ru:rw (in LocalSettings.php). I've just made it so it does, i.e. you do. See what powers you have now - David Gerard (talk) 21:32, 25 February 2013 (UTC)

I've also upgraded ru:rw from 1.19.1 to 1.19.3, you should see no difference at all but it's a little more secure - David Gerard (talk) 21:40, 25 February 2013 (UTC)

Now it works, thx!--Mr. B 22:40, 25 February 2013 (UTC)

Just broke login on ru:rw

Owing to the 1.19.3 upgrade breaking some fragile rubbish elsewhere, I just broke login on ru:rw - you can log in by going to en:rw and logging in there, while we fix it - David Gerard (talk) 21:49, 25 February 2013 (UTC)

I have just kludged it HORRIBLY, by following this kludge advice and wrapping all of /home/rationalwiki/public_html/w/includes/SkinTemplate.php in

<?php
 if ( ! class_exists ('SkinTemplate') ) {
...
}

- this is just terrible on so many levels: it fixes the problem by making it worse, it means upgrades won't quite behave and it seems to give random blank pages on login/logout. Gah. Trent's viva is next week, if we can hold on with the kludge until then that would be lovely! - David Gerard (talk) 22:13, 25 February 2013 (UTC)

As I know, the possibility of logining in and registration at ru-rw has been disabled due to spam-atacks. Registration form at ru-rw was too simple, and registration form at en-rw has been significantly supplemented and it was preferred.--Mr. B 22:40, 25 February 2013 (UTC)

... it's supposed to be broken? Are you sure? If so, I'll change it back ... - David Gerard (talk) 22:46, 25 February 2013 (UTC)

Trent decided that it would be better. At least, it's true about registration page (and maybe not about login in page).--Mr. B 00:39, 26 February 2013 (UTC)

I've changed it back as it was breaking the wiki really badly and causing all those 503s and white pages. (This bug, which is the interaction of APC and terrible code, and what I did there causes terrible code.) I can't baby it overnight, so I'll have to fix it in the morning. Or at least start on working out wtf is happening in the morning - David Gerard (talk) 00:14, 26 February 2013 (UTC)

Ok!--Mr. B 00:39, 26 February 2013 (UTC)

Wiki Love

This discussion was moved to RationalWiki:Saloon bar. Blue (is useful) 06:59, 26 February 2013 (UTC)

Help whith #if

why does :

Cet article est une ébauche{{#if:{{{1}}}| portant sur {{{1}}}{{#if:{{{2}}}|, {{{2}}}{{#if:{{{3}}}| et {{{3}}}}}}}}}

called whith Folklore as only argument return:

Cet article est une ébauche portant sur Folklore, {{{2}}} et {{{3}}}.

Jean 5 5 (talk) 20:41, 26 February 2013 (UTC)

I don't know for certain, but I think if you add a pipe (|) to each of your {{{<number>}}}'s like so: {{{2|}}}, that will fix the problem. I think I know why that works, but you'll probably need David or someone if you want an understandable explanation. Peter ^mqzp 03:25, 27 February 2013 (UTC)

Gawd, I don't understand template syntax. I wonder if the new Lua templating can be installed in 1.19 ... - David Gerard (talk) 08:28, 27 February 2013 (UTC)

Well, I appear to have fixed the problem anyway. What's so great about Lua? Peter ^mqzp 08:44, 27 February 2013 (UTC)

It's a programming language intended for programming in, rather than one where Turing-completeness was basically an accident - David Gerard (talk) 12:26, 27 February 2013 (UTC)

And where's the fun in that?

More seriously: what could we do with it that we can't do with the existing system, terrible though it may be? Peter ^mqzp 07:06, 28 February 2013 (UTC)

Well, the server isn't actually melting, so it's not in fact an immediate problem. I might not die of bracket poisoning, mostly - David Gerard (talk) 08:46, 28 February 2013 (UTC)

Thanks. Jean 5 5 (talk) 10:32, 27 February 2013 (UTC)

The reason is that the text after the pipe is included if the parameter is blank, and if you just include the pipe with no text, it includes no text. So, for instance, if the template had wibble, you'd get

Cet article est une ébauche portant sur Folklore, wibble et {{{3}}}.

...but with you get nothing at all. rpeh •T•C•E• 10:38, 27 February 2013 (UTC)

MediaWiki 1.19.4 on en:rw and ru:rw

Security patch came out a coupla hours ago, has been applied. Please let me know of any breakage (and specify new or old) - David Gerard (talk) 20:01, 4 March 2013 (UTC)

Some CSS Table on Contents thingy

A little out of my depth here but whatever.

I'm trying to limit how many levels are shown in the Table of Contents. The Goal is to get User:Voxhumana/Temp_Genesis to only show the chapter headings, and not hundreds of headings for each verse.

I think it has something to do with this (taken from here)

* 8. SOME OTHER SMALL THINGS
 * -------------------------------------------------
*/
/* Give a bit of space to the TOC */
#toc {margin: 1em 0;
}
/**
* Allow limiting of which header levels are shown in a TOC;
* <div class="toclimit-3"> , for instance, will limit to
* showing ==headings== and ===headings=== but no further.
* Used in Template:TOC
*/
.toclimit-2 .toclevel-1 ul,
.toclimit-3 .toclevel-2 ul,
.toclimit-4 .toclevel-3 ul,
.toclimit-5 .toclevel-4 ul,
.toclimit-6 .toclevel-5 ul,
.toclimit-7 .toclevel-6 ul {
display: none;

Apologies if the above is irrelevant. VOX^HUMANA 10:03, 12 March 2013 (UTC)

It seems to work, though it may take a bit of fiddling to get it looking good on the bible pages. Peter ^mqzp 23:13, 12 March 2013 (UTC)

Can you look at User:Voxhumana/Temp_Genesis and figure out what I did wrong? I keep getting the full TOC no matter what I do. VOX^HUMANA 23:18, 12 March 2013 (UTC)

You'll need to add that code to your own css page to see it, or we'll need a tech to add it for everyone. Peter ^mqzp 23:21, 12 March 2013 (UTC)

Ah. We'll need a tech then - if it only works for people who have modified their CSS that's not really a general solution. Thanks Peter VOX^HUMANA 23:36, 12 March 2013 (UTC)

I mean, you can modify your own css to test it, and once it looks good you can then get a tech to do it. Peter ^mqzp 23:38, 12 March 2013 (UTC)

Added. Look here if you need any guidance on how to use it. Looks like you've got it sorted. (formerly Ghostface Editah) 00:09, 13 March 2013 (UTC)

Sweet - MUCH appreciated. Cheers VOX^HUMANA 00:21, 13 March 2013 (UTC)

More Bible

Vox's format changes to the bible pages are screwing with the bile text generators, e.g:

Genesis 1:1-3
In the beginning God [Elohim] created the heaven and the earth.
And the earth was without form, and void; and darkness was upon the face of the deep. And the Spirit of God moved upon the face of the waters.
And God [Elohim] said, Let there be light: and there was light.

Can somebody please fix the extension? Thanks. Peter ^mqzp 00:42, 14 March 2013 (UTC)

Can you be more specific about what's wrong? (formerly Ghostface Editah) 00:47, 14 March 2013 (UTC)

There shouldn't be any '=' characters before the verses (which are themselves sometimes not showing up, but I don't know if that's related). Peter ^mqzp 00:56, 14 March 2013 (UTC)

I won't be devastated if my changes are undone in the interim - we can always restore them later. My motive is editing/reading the annotations is MUCH easier if there is a TOC to navigate. VOX^HUMANA 00:58, 14 March 2013 (UTC)

I don't see any ='s.

I have my doubts that anyone's coming to RW for bible commentary... (formerly Ghostface Editah) 00:59, 14 March 2013 (UTC)

The text doesn't appear in difflinks, but it does when you load the page/preview edits. Peter ^mqzp 01:04, 14 March 2013 (UTC)

I can see it on this page but not on the bible pages. I haven't been following what Vox did. I'll have a look and see what's going on. (formerly Ghostface Editah) 01:06, 14 March 2013 (UTC)

I have no idea how that extension works. Sorry. (formerly Ghostface Editah) 01:08, 14 March 2013 (UTC)

I don't know if anyone does, and it may be best to simply strip all uses from the site (if we can find them). How it appears to work is that the text is stored on the annotated bible pages, and is parsed by the extension to create excerpts like the one above. Changing the section headers slightly has the downside that the extension is now not parsing properly. All that probably needs to happen is for a line of code to be changed. Peter ^mqzp 01:12, 14 March 2013 (UTC)

Well I've been writing extensive Bible commentaries, and am planning to eventually get through the entire bible (I've completed 13 chapters of Genesis and a few of Joshua so far, and have offline notes on much more). However another option is to break these off into separate articles, one page per chapter. My complaint is that trying to edit a single page version is a nightmare - you are forever scrolling around trying to find where you were up to. VOX^HUMANA 02:49, 14 March 2013 (UTC)

I will look into it. Tmtoulouse (talk) 05:58, 15 March 2013 (UTC)

Fixed? Tmtoulouse (talk) 22:30, 15 March 2013 (UTC)

Looks like it. Thanks! Peter ^mqzp 01:26, 16 March 2013 (UTC)

Brilliant - thanks Trent VOX^HUMANA 01:39, 16 March 2013 (UTC)

So how about that proxy filter?

Is it turned off or something? Because our very best friend, Syamsu, has been going on a tear with various IPs, but otherwise exactly the same. Whatever we've got is insufficient against actual meaty wandals. Ochotona princeps^{not a pokémon} 10:06, 16 March 2013 (UTC)

Never mind, I wasn't paying attention to the ranges he was using. I guess he was just refreshing his IP from his ISP, not using proxies. I'd delete this, but instead, I'll just make this correction. Hurrrr. Ochotona princeps^{not a pokémon} 11:17, 16 March 2013 (UTC)

The ranges he's using are:

• 109.32.0.0/16
• 109.33.0.0/16
• 109.34.0.0/16
• 109.35.0.0/16
• 109.36.0.0/16
• 109.37.0.0/16
• 109.38.0.0/16
• 31.136.0.0/16
• 31.137.0.0/16
• 31.138.0.0/16

They're blocked for ~9 hours but I imagine he'll be back after that. rpeh •T•C•E• 11:25, 16 March 2013 (UTC)

Well that's about when I'll wake up in the morning, so I'll be around. Of course, the wife is not keen about online drinking games at 8 AM Sunday morning. VOX^HUMANA 11:29, 16 March 2013 (UTC)

What's her tipple? Make it that and it might just work. Possibly. Sophie^Wilder 11:30, 16 March 2013 (UTC)

Well, yeah: whatever we've got is insufficient against meat vandals, because if we let IPs edit to talk to us (as we do), then they can edit to trash stuff. The price of good speech being the possibility of bad speech.

The open proxy blocker is still in place. In the last days of the attack from the vandal we won't name, he was having to work really hard to find proxy IPs that weren't on the lists the blocker uses - so it's not perfect, but frankly it's good enough. ReCAPTCHA is still a really hard captcha to automate, so that was the final barrier.

At this point it's philosophical, not technical: if we want IPs to be able to talk to us, we're going to get occasional persistent morons of this sort, and it becomes a social question for the mob - David Gerard (talk) 12:18, 16 March 2013 (UTC)

Is it possible to range vandal bin?--"Shut up, Brx." 17:45, 16 March 2013 (UTC)

While block duration is topical...

Any chance of having something between 1 day and 3 months in the block duration choices? 1 week, perhaps? Sophie^Wilder 13:41, 16 March 2013 (UTC)

314159 seconds = 3.6 days - added to list - David Gerard (talk) 15:43, 16 March 2013 (UTC)

Thanks! Sophie^Wilder 16:04, 16 March 2013 (UTC)

Collapse top and bottom templates

WP has some very nifty template to collapse lengthy, irrelevant discussions that can't be justifiably deleted, but clutter up the page. There's a lengthy pile of off-topic whining going on in the Talk:Mind-Energy_forum page, and I wanted to collapse it so that the small percentage of actually relevant stuff could be responded to.

The WP templates in question are:

• Collapse top
• Collapse bottom

I would do it myself, but I'm fairly certain I'll f**k it up. VOX^HUMANA 04:57, 19 March 2013 (UTC)

IF admin rights are needed to see the code, let me know what you want retrieved and I'll get it for you. The templates look like they have admin level protection for editing, but I'm not sure if that prevents reading the code or not. VOX^HUMANA 04:59, 19 March 2013 (UTC)

We already have our own {{collapse}}, which can probably be re-purposed to fit. Of course, that doesn't mean you should: if the discussion truly is off topic (and people care enough to continue it), then moving it somewhere else might be a better option. Peter ^mqzp 06:51, 19 March 2013 (UTC)

Avoiding referers

I've created template:unref, which makes possible going around referer shenanigans such as those on a certain page starting with "aether" (intentionally unlinked). I'm just not sure how reliable it is - I guess that if they notice the links, it's possible to set the redirect to work on that specific kind of Google referers too.

Apparently, there's a client-side trick to send a page with a blank referer based on a creative use of JavaScript and "about:blank". That will require some Tech work, though.--ZooGuard (talk) 22:06, 23 March 2013 (UTC)

Link to description of said trick? - David Gerard (talk) 22:47, 23 March 2013 (UTC)

Here - it uses an iframe and a POST request. I think it can be simplified for RW's needs, as we want just the page, not POST-ing anything to it.

On RW, I imagine it to work like this: the template inserts a link that's a call to a JavaScript function, that function is in the common.js (or whatever it's called). The function opens a new tab/window from about:blank, then manipulates its contents in the way described in the linked post to refresh to or GET the target URL. Feasible?--ZooGuard (talk) 09:58, 24 March 2013 (UTC)

There used to be a http://rationalwiki.org/redir.php that I wrote because Jinx was redirecting everyone from RW to a goatse image. What happened to that? -- Nx / talk 11:23, 24 March 2013 (UTC)

May have got lost in an upgrade - David Gerard (talk) 11:41, 24 March 2013 (UTC)

Found it! Back in place at that URL. What are the parameters for it? - David Gerard (talk) 11:44, 24 March 2013 (UTC)

"?url=" seems to work.--ZooGuard (talk) 11:51, 24 March 2013 (UTC)

Send me the source please, I'd like to take a look. -- Nx / talk 12:00, 24 March 2013 (UTC)

Yeah, it doesn't work, there was a newer version, which does, but it probably has an XSS vulnerability. -- Nx / talk 12:04, 24 March 2013 (UTC)

The source is near-trivial:

<?php
$url = $_GET["url"];

header( "Location: {$url}" );
?>

- David Gerard (talk) 12:59, 24 March 2013 (UTC)

Yeah, that was my first attempt, which didn't work in all browsers (doesn't seem to work in any today) -- Nx / talk 13:44, 24 March 2013 (UTC)

Or just use anonym.to: http://anonym.to/?http://www.whatismyreferer.com/ -- Nx / talk 12:09, 24 March 2013 (UTC)

It's in a lotta workplace censorware lists - David Gerard (talk) 13:00, 24 March 2013 (UTC)

But yes, http://rationalwiki.org/redir.php?url=http://whatismyreferer.com seems to do the right thing - David Gerard (talk) 13:02, 24 March 2013 (UTC)

No, it doesn't - after clicking on your link, it pointed out that the referer's this page.

I've put anonym.to in Template:Unref for now. Feel free to change it to redir.php if/when you fix it.--ZooGuard (talk) 13:10, 24 March 2013 (UTC)

Well, c*ck - David Gerard (talk) 13:20, 24 March 2013 (UTC)