User talk:Nx/Archive14

This is an archive
Please do not edit this page
To leave a message, return to my talk page

Archives:

1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15

, (new)(back)

nav portals

Hey, thanks for fixing all those! This is what I built from them: template:portal portal. ħuman 17:54, 3 September 2010 (UTC)

*sigh*

you're so... conventional. --151.81.24.147 (talk) 17:42, 8 September 2010 (UTC)

Yeah, get those damn kids off my lawn!!! I mean your lawn. I mean wiki. --151.82.15.44 (talk) 17:07, 10 September 2010 (UTC)

Yeah, it was funny the first time, but you are starting to annoy me. -- Nx / talk 17:11, 10 September 2010 (UTC)

Damn kids have no respect these days. --151.82.15.44 (talk) 17:22, 10 September 2010 (UTC)

Portal icons

Thanks for fixing the film icon. You might want to take a look at the one on Human's new Media, too... Blue (is useful) 07:11, 12 September 2010 (UTC)

Nice try

Ok, that one is logical. The fixation that people have with the irrelevant ones is beyond me. Let's see if sanity prevails.--Bob^{Spring is sprung!} 15:20, 17 September 2010 (UTC)

Nope. -- Nx / talk 08:33, 18 September 2010 (UTC)

Stop doing mean things with my edits

I don't want to undo my edit at least not until you told me in my talk page. So if anyone of us wants to undo somebody edit they must tell the other person before doing it. If you find one of edits and say I want to delete it tell me first. ^{— Unsigned, by: Foliver / talk / contribs} 21:04, 30 September 2010 (UTC)

First of all, Please sign talk page entries using four tildes like this: ~~~~ or by clicking on the sign button: , on the toolbar above the edit panel. Thank you.

Secondly, So if anyone of us wants to undo somebody edit they must tell the other person before doing it no, that's not how it works. -- Nx / talk 21:08, 30 September 2010 (UTC)

Your bridge

was taxpayer funded and as such belongs to We the People, not to you. Get thee to an Obamaville. EddyP (talk) 18:59, 2 October 2010 (UTC)

How dare you waste MY taxpayer money on such a nonsensical bridge. There's not even any water under it! Lord Goonie ^{Hooray! I'm helping!} 19:01, 2 October 2010 (UTC)

Where's that confounded bridge? ħuman 02:28, 3 October 2010 (UTC)

That's not the talk page

Thanks for fixing that, us noobs always need help! ħuman 08:29, 10 October 2010 (UTC)

Why are you doing that?

Willminator doesn't deserve that and he didn't ask for it. 04:46, 13 October 2010 (UTC)

It contains JeffT's real name and another real name in one edit. It doesn't matter who deserves it and who doesn't. -- Nx / talk 04:48, 13 October 2010 (UTC)

I doubt the names are "real". A simple rollback would have been sufficient. ħuman 04:55, 13 October 2010 (UTC)

We don't know if they're real but it doesn't matter, those are the rules. -- Nx / talk 05:19, 13 October 2010 (UTC)

Look Sam, I know your real last name is Nigelsson! ħuman 05:28, 13 October 2010 (UTC)

Help

Nx, t'other day you helped me get my Hitachi drive mounted, well it's gone again!

screenshot

Oh and now it's doing the same with my SD card. HALP!!! 21:04, 22 October 2010 (UTC)

Unplug replug didn't work? The system is being stupid and it's trying to mount your internal drive, which is obviously already mounted. I suspect the drive doesn't get enough power during bootup and it borks, hence my suggestion to un/replug. Or better yet don't plug it in before boot. The mount command I gave you should still work though. -- Nx / talk 21:19, 22 October 2010 (UTC)

Seems to have sorted without any intervention on my part. Thanks anyway Nx. 19:44, 25 October 2010 (UTC)

Regarding pywikipedia

Any idea why when I ran category.py it only worked on a few pages in Category:US Supreme Court Justices? Blue (is useful) 22:30, 23 October 2010 (UTC)

Took three times to work fully. Blue (is useful) 22:50, 23 October 2010 (UTC)

Update: I has fixed it. Blue (is useful) 22:58, 23 October 2010 (UTC)

Actually, new problem. Keep getting "internal service 500 error" or thereabouts, and "can't access rationalwiki/api.php" or thereabouts. ??? Blue (is useful) 05:26, 24 October 2010 (UTC)

I use an old version of pwb because newer versions didn't work at all for me. It's rationalwiki.org/wiki/api.php btw -- Nx / talk 05:49, 24 October 2010 (UTC)

Hmm. It seems from a preliminary analysis that the portal templates that "includeonly" a category are screwing with category.py. Blue (is useful) 19:33, 24 October 2010 (UTC)

Here's my version of pwb: [1] -- Nx / talk 19:45, 24 October 2010 (UTC)

That made it completely unusable. Blue (is useful) 19:16, 25 October 2010 (UTC)

But thanks anyway. Blue (pester) 19:17, 25 October 2010 (UTC)

I didn't include the login-data directory (since that contains nxbot's login data), maybe that's the problem? -- Nx / talk 19:22, 25 October 2010 (UTC)

I FIGURES IT OUT! It's the dpl. Tests confirm. For some reason pwb can't connect to rationalwiki.org/wiki/api.php when any templates with dpl tags are transcluded. Blue (pester) 19:26, 25 October 2010 (UTC)

Oh goody. Let me check the server logs. -- Nx / talk 19:27, 25 October 2010 (UTC)

[25-Oct-2010 15:19:13] PHP Fatal error: Call to a member function getNamespace() on a non-object in /home/rational/public_html/wiki/extensions/DynamicPageList/DPLMain.php on line 2665 -- Nx / talk 19:28, 25 October 2010 (UTC)

So that's... good? Blue (pester) 19:32, 25 October 2010 (UTC)

Fatal is usually not good. -- Nx / talk 19:33, 25 October 2010 (UTC)

Heh, no, I mean good that the problem has been pinpointed. Is it resolveable? Blue (is useful) 19:35, 25 October 2010 (UTC)

Yes, it's good in one sense, bad in the sense that I have to fix it.

But joking aside, thanks for finding the cause, I do appreciate it. -- Nx / talk 19:39, 25 October 2010 (UTC)

No problem, I'm just glad to get to use my bot for these menial tasks. Blue (pester) 19:41, 25 October 2010 (UTC)

You up for some brain storming?

I am looking to move RW back to commercial hosting, and have been trying to test a migration over to a new server. I have things setup over at |beta for now. I am getting internal errors on extensions when trying to edit. Could that be explained as using the config files from rw.org on rwbeta.com, or is there likely something more complicated and sinister going on? tmtoulouse 18:51, 30 October 2010 (UTC)

That's probably from liquidthreads. I think I left rationalbeta in a half-broken state last time I updated lqt, and couldn't be arsed to fix it. Disable the extension and it should work. -- Nx / talk 18:53, 30 October 2010 (UTC)

Actually rationalbeta now points to the new server with a copy of the current installation on my server, and that's whats failing. They always make backups, restores, and exports sound so easy....tmtoulouse 18:54, 30 October 2010 (UTC)

Oh, it's a copy of RW, not RWBeta. The problem is that the recaptcha key is wrong in localsettings. Look at the old rationalbeta's config, find the recaptcha keys, and change them to that. Strange that it aborts like that though. The ones in RW's config are only valid for the domain name rationalwiki.org, that's why it doesn't work. -- Nx / talk 18:59, 30 October 2010 (UTC)

Actually, recaptcha doesn't list rationalbeta.com, how odd. Eh, just create a new key for rationalbeta.com with your own recaptcha account. -- Nx / talk 19:03, 30 October 2010 (UTC)

For testing I just disabled recaptcha, then it complained about vandalbreak, disabled that and it moved on to the next extension. So its something bigger. Doing some poking around apparently php 5.3 is buggy with mediawiki going to try and downgrade and see if that works. tmtoulouse 19:04, 30 October 2010 (UTC)

And yep that was the problem. Sorry to bug you my google fu was just delayed. tmtoulouse 19:12, 30 October 2010 (UTC)

How the hell did you downgrade php on ubuntu that fast? Anyway, if we're going to do a server move, we might as well upgrade to MW1.16 while we're at it, that should be compatible with php 5.3.2 (there's a bug in 5.3.1). And can we have a 64-bit OS or is that too much work? -- Nx / talk 19:21, 30 October 2010 (UTC)

I used the repositories for an older version of ubuntu and just did an apt-get install on php 5.2 for each package. Anyway, yeah the server has the latest version of ubuntu 64 bit installed on it right now. Is dpl working with the new mw version? tmtoulouse 19:26, 30 October 2010 (UTC)

Mixing packages like that is probably not a great idea... Sadly, DPL is kinda stagnant, I wish we could just get rid of it, but we can't. Anyway, I can try fixing the bug myself. What's the deadline? -- Nx / talk 19:35, 30 October 2010 (UTC)

Wouldn't it be more prudent to separate the move and the upgrade so they can be debugged individually? Or is that irrelevant? ħuman 19:36, 30 October 2010 (UTC)

Well the "move" is just pointing rw.org to the new server whenever it appears to be working. So updating MW is just one more step to do before changing the DNS. I would be inclined to go for it, but I have held of really pushing for an MW update because I didn't think DPL was working and we really use it. tmtoulouse 19:42, 30 October 2010 (UTC)

Yeah, between the DPL problems, LQT not being ready and me backporting a few of the features I cared about in MW 1.16, there's not really a huge incentive, but I don't really want to hold it off either, because 1.14 is already obsolete and not supported any more. -- Nx / talk 19:45, 30 October 2010 (UTC)

Well there is no inherent deadline, I would like to move to the new server before I leave for the break, which is the second week of dec. We also don't have to have the software upgraded when we move. The service I am working with lets you easily clone a VPS for any amount of time which would make testing updates a lot easier. tmtoulouse 19:40, 30 October 2010 (UTC)

Isn't a virtual server going to cause problems when our CPU usage skyrockets, like when you decided to move away from commercial hosting? -- Nx / talk 19:45, 30 October 2010 (UTC)

They use a weighted CPU allocation for when the server is under load, but other wise free cycles are up for grabs by anyone. The CPU issue was a problem when we were using cheap shared hosting. VPS worked great for us but was too expense at the time. Slicehost is cheaper than what we were using, and there is more money to throw at the site now then there was before. tmtoulouse 19:51, 30 October 2010 (UTC)

Though a quick question regarding the current computer infrastructure; where is it going? I understand you transferred ownership of it to the RWF. Would you keep if for RW related activities like backups or hosting other related projects or do you just want it back? ^bomination 22:03, 30 October 2010 (UTC)

The foundation still owns it, likely continue to use atleast some of it for running automated scripts and what not, such as capturebot. tmtoulouse 22:08, 30 October 2010 (UTC)

Honestly, for all intents and purposes I think the new server is "ready." There appears to be some sort of mysql query that is locking up I am trying to figure out whats going on there, and I don't think php-apc is working. Once I get those straightened out might as well start planning the move, that is unless you think its worth trying to pound out a MW upgrade. I haven't even looked at the DPL bug issue though. tmtoulouse 20:35, 31 October 2010 (UTC)

I'll need some time to get things fixed for MW1.16 (I didn't do much testing with RationalBeta, but things seemed to work nicely). But we can't prepare for everything, so we'll just fix stuff as it comes up. I'm not really comfortable with the downgrade you did with php, that kind of stuff is a recipe for disaster on Ubuntu. When do you have time to do this? -- Nx / talk 07:26, 1 November 2010 (UTC)

Well, basically this week, then not really much at all till December. With the goal of getting the switch before the 16th or 17th. If you are up for putting in some leg work on this (and the help is always appreciated) I can hook you up with access to the new server.

MYSQL is still hanging randomly, though I did a myisamchq repair on the tables and it hasn't happened since, but we shall see. Things seem a little buggy because I transferred the database from a mysqldump copy, but some of the tables only have a .frm file and so the copy doesn't work. I have to pull those in using a the .sql file that generates the tables. So that seems a little sketchy to me as well. But it "works."

As for the php downgrade this is essentially what I ran:

sudo apt-get install php5=5.2.10.dfsg.1-2ubuntu6 php5-dev=5.2.10.dfsg.1-2ubuntu6 php-pear=5.2.10.dfsg.1-2ubuntu6 php5-cgi=5.2.10.dfsg.1-2ubuntu6 php5-cli=5.2.10.dfsg.1-2ubuntu6 php5-common=5.2.10.dfsg.1-2ubuntu6 php5-mysql=5.2.10.dfsg.1-2ubuntu6

Is it problematic? tmtoulouse 07:34, 1 November 2010 (UTC)

I don't know if that command locks the packages to that version or if the next time you run apt-get upgrade it will update them back to php5.3. But it's definitely not an ideal solution. I'll see what I can do about the Dpl bug, if I can fix it (either today or probably tomorrow) then we can do it this week. -- Nx / talk 08:49, 2 November 2010 (UTC)

Playing around with php-apc I wasn't explicit with my version numbers and it updated to php 5.3 and I had to downgrade again. So its itching to go back to 5.3 but editing seems to be impossible with that version. However, I seem to have things totally broken at the moment anyway. And I can't figure out why the database is hanging. So their is enough wrong at the moment that there is probably no hurry after all. tmtoulouse 17:08, 3 November 2010 (UTC)

An excuse to be lazy? You don't have to tell me twice. -- Nx / talk 17:14, 3 November 2010 (UTC)

Forum tech question

Why isn't Forum:2010_"World"_Series turning up in the recently-edited list at the SB? Did I do something wrong when I created it? ħuman 22:22, 30 October 2010 (UTC)

fixed it. -- Nx / talk 07:20, 31 October 2010 (UTC)

Thank you very much! ħuman 09:58, 31 October 2010 (UTC)

More questions

In my autosigning bot, everything goes swimmingly until it finally tries to update the page -- and I get an edit conflict every time. Thoughts/solutions? Blue (is useful) 22:14, 31 October 2010 (UTC)

That code you posted looks weird, you should do it like this:

site = wikipedia.getSite('en','rationalwiki')
site.forceLogin()
page = wikipedia.Page(site,pagename)
page.put(text,summary,False,False) #the two bools are watchArticle and minorEdit

This works for capturebot. Anyway, you will have to prepare your bot for edit conflicts, imagine autosigning a busy page like the saloon bar. -- Nx / talk 07:37, 1 November 2010 (UTC)

Btw, capturebot's code is here -- Nx / talk 07:50, 1 November 2010 (UTC)

So, I actually think my bot is ready for beta testing, but it still can't update pages with DPLs, which is problem for the SB (just the bar? I think so). Any chance you've figured out the problem? Blue (pester) 03:07, 4 November 2010 (UTC)

Actually, your talk page uses dpls in some way. My stunning example has failed. Blue (pester) 03:18, 4 November 2010 (UTC)

For testing purposes, I'll have to specifically tell the bot not to monitor any page with dpls, which would exempt the SB, your talk page, and others(?). Blue (is useful) 03:58, 4 November 2010 (UTC)

Your talk page (and the SB) is now immune to autosigning. Congratulations! Blue (is useful) 04:08, 4 November 2010 (UTC)

Can you try editarticle.py on my talk page? It works for me. -- Nx / talk 06:07, 4 November 2010 (UTC)

Anyway, I think I know what the problem is and how to fix it, but need your help since I can't reproduce the bug apparently. -- Nx / talk 06:17, 4 November 2010 (UTC)

Ok, I fixed it since it seems stubbot was triggering it too, should be working now. -- Nx / talk 06:27, 4 November 2010 (UTC)

Trying again to edit using editarticle.py. Blue2 (talk) 06:27, 4 November 2010 (UTC)

I think it worked.

Well, where's the sig? Anyway, you should unbot your bot for the test run so we can notice if it screws up. -- Nx / talk 06:33, 4 November 2010 (UTC)

(EC) Forgot to tell the bot to try to edit your talk page. Blue (pester) 06:35, 4 November 2010 (UTC)

Well? -- Nx / talk 06:44, 4 November 2010 (UTC)

I'm trying to figure out what happened. Blue (is useful) 06:50, 4 November 2010 (UTC)

It's a very fickle script ^{— Unsigned, by: Blue / talk / contribs at 06:56, 04 November 2010 (UTC)}

Some times it works, sometimes it doesn't. ^{— Unsigned, by: Blue / talk / contribs at 07:05, 04 November 2010 (UTC)}

bugs bugs bugs Blue (pester) 07:07, 4 November 2010 (UTC)

If you'd rather be fixing dpl, just say so :D -- Nx / talk 07:09, 4 November 2010 (UTC)

As an aside, which one (or none) are you? Blue (is useful) 15:51, 5 November 2010 (UTC)

Wikipedia:Starship registry (Star Trek)#NX -- Nx / talk 16:55, 5 November 2010 (UTC)

Enlightening. And as another aside, is there a way using pwb to generate a list of a user's contributions? (like, of course, special:contributions) Blue (is useful) 20:02, 5 November 2010 (UTC)

check userlib.py -- Nx / talk 20:07, 5 November 2010 (UTC)

multas gratias tibi ago. Blue (is useful) 20:23, 5 November 2010 (UTC)

Another question: how does one generate the users in the creation log? Blue (pester) 20:58, 5 November 2010 (UTC)

Let me clarify: using pyb, not the HTML way? Blue (pester) 03:53, 6 November 2010 (UTC)

There's a welcoming bot in pwb (welcome.py or something like that), your answer should be there. -- Nx / talk 06:32, 6 November 2010 (UTC)

Thanks again. How's the autosign bot? Blue (is useful) 06:34, 6 November 2010 (UTC)

Was too tired to work on it yesterday evening, and now the server move is a priority. I'll try to get some work done on it afternoon -- Nx / talk 06:42, 6 November 2010 (UTC)

How to upgrade?

I am having a hard time imaging the ideal plan of action for the server move+upgrade. If we upgrade everything on the new server doesn't that mean rsyncing the database from RW when we move could really fuck things up? It almost seems like we have to move to the new server first. Tmtoulouse (talk) 17:56, 5 November 2010 (UTC)

The way I imagine it is: copy over a database dump and import it, copy over the upload directory, set up a clean MediaWiki installation (we've got a lot of leftover cruft), install all our extensions, configuration files and other stuff, then run the upgrade script. When I upgraded RationalWikiWiki, I essentially did the same thing, minus having to move the database of course. -- Nx / talk 18:12, 5 November 2010 (UTC)

But all this would have to be done post-move right? Or we would have to lock the database here cause once we upgrade on the server we can't bring this old database back over. So if we want to do a proof of concept upgrade on the new server we would have to do it, test it, then revert back, move over current database/uploads and start again. Or we could move to the new server with current versions, and do the proof of concept back on this box before doing it on the new server. Short version: shouldn't we move to the new server first? Tmtoulouse (talk) 18:21, 5 November 2010 (UTC)

I think it would be best to do the proof of concept there, it would mean less downtime for the site. Plus things would be cleaner that way, if we just moved over we'd have the php problem too. Once we iron out things (I've already done some of that at old rationalbeta) we can just lock the db and do it again, and if anything goes wrong we can fall back easily. Ideally all people would notice is the db lock, then when the dns propagates they'll be at the new site. -- Nx / talk 18:43, 5 November 2010 (UTC)

Allright, I unborked the database on the new server finally. But the upload directory I think has lagged. So which version are we upgrading too? Tmtoulouse (talk) 18:47, 5 November 2010 (UTC)

I think we should install 1.16.0 first (just to be safe) then immediatelly upgrade to the branch wp i using. Do you want to start doing this nowish? i.e. should I get up and turn my computer on? -- Nx / talk 18:54, 5 November 2010 (UTC)

Well I am trying to undue all my tinkering to get the old php installed and reupgrade everything then I need to eat and get caffeine or I will break something. But yeah I would like to start poking at it. Do you want me to setup a sudo account on the new server for you?Tmtoulouse (talk) 19:11, 5 November 2010 (UTC)

Yeah, do I can't promise to do a lot on the weekend. If it rains on Sunday, then I'll have the entire day. -- Nx / talk 19:25, 5 November 2010 (UTC)

When we lock the db, we should do it during off-peak hours. Preferably at night. The mechanism should be something we can easily turn off. Like a mode. We could call it "night mode". Occasionaluse (talk) 19:19, 5 November 2010 (UTC)

<------------------Saloon bar is that way, happy trolling. Tmtoulouse (talk) 19:21, 5 November 2010 (UTC)

Not even a little, huh? This wiki sure is SRS BZNS... Occasionaluse (talk) 19:38, 5 November 2010 (UTC)

I am under horrid time constraints! I can only justify feeding the troll when the trolling is new! Tmtoulouse (talk) 20:24, 5 November 2010 (UTC)

Understood...proceed with the productivity. I still think a throwback is nice. Occasionaluse (talk) 21:17, 5 November 2010 (UTC)

The mess I made

Okay you should have an e-mail with details about a user account, let me know if you don't get it. I have upgraded to the latest version I believe. The following extensions had to be disabled to allow editing at all:

• recaptcha
• vandalbrake
• bestof
• gadgets

Beyond that I haven't tested other extensions. Also special pages is completely borked. That is all I have at the moment. Tmtoulouse (talk) 21:10, 5 November 2010 (UTC)

Just for reference I have started a notes over on User:Tmtoulouse/upgrade. Tmtoulouse (talk) 21:35, 5 November 2010 (UTC)

I got the email, thanks, but I'm too tired now, sorry. I think I know what's wrong with those extensions, if it's what I think I've already fixed it on the old rationalbeta. -- Nx / talk 22:12, 5 November 2010 (UTC)

No worries, nothing has to be done now. I will keep poking here and there and make notes about what I find. Tmtoulouse (talk) 22:13, 5 November 2010 (UTC)

Pulling user group in an extension

If I query for user groups in an extension does it only do it for someone that "saves" the extension to the page, or does it happen when a user loads a page with the extension up. I am playing with ideas for collecting votes, and created an autopromote group based on criteria for voting. I query it in the extension and was just trying to output a "you are eligible' message or not. But it seems hit or miss. I didn't know if this was a cache issue or further up the chain. Tmtoulouse (talk) 17:31, 6 November 2010 (UTC)

$wgUser should be the person viewing the the page. It shouldn't be affected by caching -- Nx / talk 17:36, 6 November 2010 (UTC)

I was trying to grab it using "$groups = $wgUser->getEffectiveGroups();" but if its not affected by caching then I broke something else. Thanks! Tmtoulouse (talk) 17:41, 6 November 2010 (UTC)

I may be wrong though, try echoing the user name or something to make sure it's the right user. -- Nx / talk 17:45, 6 November 2010 (UTC)

Eh, I think its working when I pass it as a hidden value in a form, I am trying to pass it back to the extension but now I am confused. I will keep poking at it.

On another note, the randomselection extension appears borked on beta but the parser function #choice works fine. Is there any reason I should not just make a general call for people to start converting anything using the <option> to the parser function? Tmtoulouse (talk) 22:51, 6 November 2010 (UTC)

Give me a chance to fix it first, but yeah, the parser function version is generally better. -- Nx / talk 06:20, 7 November 2010 (UTC)

Sure, just trying to think of ways to reduce workload. Tmtoulouse (talk) 06:22, 7 November 2010 (UTC)

Having to replace all the choice tags would increase it I think. -- Nx / talk 06:27, 7 November 2010 (UTC)

If you and I had to do it, but if I said that everyone else had to do it....Tmtoulouse (talk) 06:28, 7 November 2010 (UTC)

I'd end up having to fix various screwups (putting | in front of the first option, subtle problems due to this being a parserfunction not a tag, etc.) -- Nx / talk 06:36, 7 November 2010 (UTC)

Spambots

Maybe we should install captcha, either for signing up or for external links? Or is it not that much of a problem? Totnesmartin (talk) 11:16, 7 November 2010 (UTC)

We do have it. You are giving the spammer too much credit thinking it is a bot, you can pay a guy to spam your site for you for $0.30 an hour. - π 11:18, 7 November 2010 (UTC)

I didn'r know we had captcha...haven't created a sock in ages... Totnesmartin (talk) 16:26, 7 November 2010 (UTC)

Capturebot

seems borked --Onion ^Hi! :) 01:03, 9 November 2010 (UTC)

He needs to update the python family file thing. I'll have a quick peek. - π 01:35, 9 November 2010 (UTC)

Nx == idiot. Should be fixed now. Pi, do you have the botserver's IP address? I don't want to make the bot report its IP address on wiki because they'd block it from viewing CP. -- Nx / talk 06:13, 9 November 2010 (UTC)

I think the IP address has changed again, I kept getting a time out error today. - π 11:13, 9 November 2010 (UTC)

I'll email you. -- Nx / talk 11:28, 9 November 2010 (UTC)

Thanks. - π 12:58, 9 November 2010 (UTC)