[00:09:57] <neilk_>	ok, back in channel... sumanah, anyone else want to talk about an API sandbox?
[00:10:26] <Reedy>	I suppose as I'm here, I should listen in ;)
[00:10:40] <neilk_>	BTW, the Flickr API explorer is not a sandbox, it's just a web interface that helps you construct real API calls, and see what the result will be.
[00:10:47] <neilk_>	Flickr generally tries to do things with maximum danger
[00:10:58] <Reedy>	Something AJAXy then?
[00:10:59] <apergos>	:-D
[00:11:12] <^demon|away>	For maximum danger we should allow editing the MediaWiki source via the MediaWiki namespace.
[00:11:41] <apergos>	I think having people test on test.wp when someone could sync their test right out to the cluster is...
[00:11:44] <apergos>	close enough
[00:11:44] <neilk_>	Reedy: here's a sample. http://www.flickr.com/services/api/explore/?method=flickr.photos.search
[00:12:11] <neilk_>	Reedy: in order to make that work, there is just metadata that describes what the arguments are, and whether they are optional
[00:12:26] <Reedy>	Cool
[00:12:31] <neilk_>	Reedy: and some sample data that would work for you, on the right hand side
[00:12:44] <Reedy>	We can do that programatically now, as the api exposes "requiredness" through paraminfo
[00:13:14] <Reedy>	Or not
[00:13:15] <Reedy>	Seemingly
[00:13:16] <neilk_>	We already have that information in our API definition, actually better than Flickr did it (there's a separate thing you have to edit to make the API explore page work)
[00:13:16] <Reedy>	I wonder
[00:13:47] <neilk_>	I have a mw.Api module that could make constructing the call pretty easy, too
[00:13:57] <neilk_>	and getting permission
[00:13:58] <Reedy>	<param name="repo" description="Name of repository to look at" required="" type="string" />
[00:14:27] <neilk_>	anyway did we post this as a Gsoc project or something?
[00:14:29] <neilk_>	It's a good idea
[00:14:51] <Reedy>	"API sandbox environment (a la Flickr API Explorer that gives a form interface to help learn the api [1])"
[00:14:53] <Reedy>	yeah
[00:14:59] <neilk_>	rather than call it API sandbox, it should just be an API constructo-thingy, and we'll make a sandbox wiki for it.
[00:15:10] <Reedy>	Mmm
[00:15:16] <bawolff>	We could just make the api output it
[00:15:20] <Reedy>	heh
[00:15:24] <bawolff>	then people could play with their local wiki
[00:15:30] <Reedy>	An extension to do it via special page would be cool
[00:15:39] <neilk_>	this looks totally easy, IMO
[00:15:40] <bawolff>	Its kind of like a fancy version of the help module
[00:16:54] <bawolff>	and it would really help. I have fun memories of writing ajax functions in js just to test the editing interface/other stuff requiring being posted (read: was a pain)
[00:18:08] <Reedy>	hmmm
[00:19:22] <^demon>	Nobody's going to do the project I suggested because it's kind of boring :(
[00:19:37] <Reedy>	{{sofixit}}
[00:19:38] <Reedy>	:D
[00:19:56] <^demon>	I should add some buzzwords to it.
[00:20:10] <bawolff>	lol
[00:20:18] <Reedy>	vandal stopping!
[00:20:35] <^demon>	"Rewrite blocking to be more semantic leveraging synergy between HTML5 and web 2.0"
[00:20:51] <apergos>	ow ow
[00:20:55] <apergos>	that's painful even to read
[00:21:06] *Reedy 	holds a baseball bat near ^demons head...
[00:21:43] <neilk_>	^demon: what is "blocking" code?
[00:21:50] <neilk_>	you mean the parser?
[00:21:58] <^demon>	No, the code relating to user blocks.
[00:22:02] <^demon>	In the Block class :)
[00:22:02] <neilk_>	oh
[00:22:02] <Reedy>			$form = new IPBlockForm( '' );
[00:22:02] <Reedy>			$form->BlockAddress = $params['user'];
[00:22:08] <neilk_>	right
[00:22:12] <Reedy>	So the API doesn't have to do shit like ^
[00:22:13] <neilk_>	is that even IPv6 compatible?
[00:22:20] <Reedy>	Ask Aaron
[00:23:05] <^demon>	Mostly, I think.
[00:23:16] <Reedy>	7 ipv6 outstanding
[00:23:18] <^demon>	I think the major outstanding problem was range blocks in ipv6.
[00:23:21] <Reedy>	and not all directly MW
[00:23:57] <Reedy>	"Should block IPv6 addresses at /64 instead of /128"
[00:23:58] <Reedy>	https://bugzilla.wikimedia.org/buglist.cgi?keywords=ipv6&resolution=---
[00:53:41] <neilk_>	reedy: i was totally doing that
[00:53:45] <Reedy>	:D
[00:53:47] <neilk_>	reedy: but thanks anyway
[03:17:31] <Shirley>	So this sumanah character, does he have an e-mail address or something?
[03:17:37] <Shirley>	robla: ^
[03:17:47] <jorm>	she.
[03:17:57] <Shirley>	Really? All right.
[03:18:13] <Shirley>	I had to deal with a "Morgan" today at work. Ended up excluding the prefix altogether.
[03:19:15] <Shirley>	I'd like to stress to her that GSOC projects should be manageable. No more LiquidThreads and interwiki transclusions and other crazy things that require a really indepth knowledge of MediaWiki.
[03:19:30] <Shirley>	(And end up not producing anything.)
[04:24:48] <robla>	Shirley: sumanah's email address is ??sumanah@panix.com, but I fully agree with you anyway
[04:25:43] <robla>	She's going to be doing more facilitation than picking the projects (similar to the role I played last year)
[04:26:35] <robla>	So....I'm probably more important to make that case to (as someone who is planning to be more vocal+forceful about project selection)
[11:03:49] <apergos>	Fatal error: Call to protected method  LanguageConverter::translate() from context 'AbstractFilter' in  /usr/local/apache/common-local/php-1.17/extensions/ActiveAbstract/AbstractFilter.php  on line 123
[11:04:12] <apergos>	time to see when that got protected and what the preferred replacement is and nag someone
[11:40:32] <TimStarling>	maybe convert()?
[12:59:45] <apergos>	how can I specificy a specific variant to convert()?
[13:00:13] <apergos>	abstracts needs output in both zh-cn and zh-tw
[13:00:17] <apergos>	( TimStarling )
[13:01:21] <TimStarling>	you probably want autoConvert(), you can specify the variant in the second parameter
[13:03:29] <apergos>	it seems to do a lot more work
[13:05:56] <TimStarling>	you could just make translate() public
[13:06:08] <apergos>	that was my plan actually
[13:06:40] <TimStarling>	http://www.mediawiki.org/wiki/Special:Code/MediaWiki/69133
[13:06:54] <TimStarling>	no explanation
[13:06:58] <apergos>	yeah I had found it earlier
[13:07:15] <apergos>	not a lot of commentary no
[13:07:28] <apergos>	I'll comment when I change it back
[17:56:43] <RoanKattouw>	TrevorParscal: 10:02 am, reporting for duty ;)
[18:06:26] <TrevorParscal>	howdy
[18:06:27] <TrevorParscal>	ok
[18:07:02] <RoanKattouw>	You wanted to deploy AF to prototype
[18:07:12] <TrevorParscal>	so, I just need the latest ArticleFeedback in trunk to be merged into wherever the prototypes are pulling from
[18:07:27] <TrevorParscal>	then the prototype to be updated
[18:07:34] <TrevorParscal>	and then the config to be changed
[18:07:37] <RoanKattouw>	OK
[18:07:45] <RoanKattouw>	I take it you wanted to use Priyanka's existing prototype setup?
[18:07:58] <RoanKattouw>	And Priyanka's down under now, right?
[18:07:59] <TrevorParscal>	yeah, that's what's in place now, we should roll with that
[18:08:02] <TrevorParscal>	yeah
[18:09:30] <RoanKattouw>	URL: http://svn.wikimedia.org/svnroot/mediawiki/branches/wmf/1.17wmf1
[18:09:51] <RoanKattouw>	So that's where it's pulling from
[18:10:06] <RoanKattouw>	I'm just gonna check out ArticleFeedback in the extensions directory locally
[18:10:26] <RoanKattouw>	So I don't have to mess around with the wmf branch just yet
[18:10:47] <RoanKattouw>	catrope@prototype:/srv/org/wikimedia/prototype/wikis/rc/extensions$ sudo svn checkout http://svn.wikimedia.org/svnroot/mediawiki/trunk/extensions/ArticleFeedback
[18:11:04] <RoanKattouw>	Damn that checkout was FAST
[18:11:08] *RoanKattouw 	<3 formey
[18:12:38] <RoanKattouw>	TrevorParscal: OK I left a require_once() statement in the config file, I'
[18:12:44] <RoanKattouw>	TrevorParscal: I'll leave it to you to put the config in
[18:13:06] <RoanKattouw>	 /srv/org/wikimedia/prototype/wikis/rccommon/PrototypeSettings.php line 138
[18:15:06] <^demon>	RoanKattouw: It's less to do with formey and more than it's in the same datacenter as prototype :p
[18:15:23] <RoanKattouw>	I know that
[18:15:28] <RoanKattouw>	That's exactly what I like about formey
[18:15:32] <^demon>	:)
[18:15:39] <TrevorParscal>	ok
[18:15:40] <^demon>	makes fenari -> formey way fast too
[18:15:55] <RoanKattouw>	Yeah. When I ran make-wmf-branch, SVN from my home was so slow that it was faster to compile SVN 1.6 on fenari and run the script there
[18:16:06] <RoanKattouw>	(I raced the two approaches)
[18:16:34] <TrevorParscal>	hmmm http://prototype.wikimedia.org/release-en/AricleFeedbackTest
[18:16:46] <TrevorParscal>	I added the settings, but I'm still seeing the old version
[18:17:26] <RoanKattouw>	I wonder if ArticleAssessmentPilot is still enabled
[18:17:32] <TrevorParscal>	not that i can tell
[18:17:48] <RoanKattouw>	It is
[18:17:52] <RoanKattouw>	On Special:Version
[18:18:51] <RoanKattouw>	It was enabled in CommonSettings.php , I disabled it
[18:21:50] <RoanKattouw>	Is $wgArticleFeedbackOdds a number in 0-10 ?
[18:22:11] <TrevorParscal>	0-100
[18:22:22] <TrevorParscal>	hmm
[18:22:25] <TrevorParscal>	this is odd
[18:22:31] <TrevorParscal>	it's not showing up...
[18:22:37] <RoanKattouw>	Oh haha
[18:22:38] <RoanKattouw>	Got it
[18:22:40] <RoanKattouw>	>>> mw.config.get('wgArticleFeedbackCategories',[])
[18:22:43] <RoanKattouw>	["ArticleFeedback"]
[18:23:01] <TrevorParscal>	I fixed that in the config just a sec ago
[18:23:07] <TrevorParscal>	to have a space
[18:23:32] <RoanKattouw>	It's in the startup module, so it sticks
[18:23:50] <RoanKattouw>	OK got it now
[18:23:55] <TrevorParscal>	yeah, me too
[18:23:59] <TrevorParscal>	the tool?
[18:24:02] <TrevorParscal>	or the value?
[18:24:09] <RoanKattouw>	The value
[18:24:15] <TrevorParscal>	right, me too
[18:24:16] <RoanKattouw>	var wonLottery=(Number(mw.config.get('wgArticleId',0))%1000)
[18:24:18] <RoanKattouw>	28<Number(mw.config.get('wgArticleFeedbackLotteryOdds',0))*100;
[18:24:34] <RoanKattouw>	To me that looks like the odds are in 0-10
[18:24:51] <TrevorParscal>	hmm
[18:24:52] <RoanKattouw>	Because if odds == 10, the right hand side will be 1000, so wonLottery will always be true
[18:25:00] <RoanKattouw>	Doesn't matter though, odds=50 right now
[18:25:08] <RoanKattouw>	So the chances of it showing should be 500% :D
[18:25:09] <TrevorParscal>	crap, I got that all backwards i guess
[18:25:11] <TrevorParscal>	but still
[18:25:12] <TrevorParscal>	yeah
[18:25:26] <TrevorParscal>	i have the odds at 100 actually
[18:25:31] <TrevorParscal>	but that aside
[18:25:42] <TrevorParscal>	the category should make it show up
[18:26:39] <RoanKattouw>	Hm the first thing I notice in ext.articleFeedback is another lottery function
[18:27:34] <TrevorParscal>	thats for pitches only
[18:28:23] <RoanKattouw>	Oh
[18:28:37] <RoanKattouw>	Hmph why won't Firebug let me put breakpoints in this script
[18:28:41] <RoanKattouw>	It's not showing in the script panel
[18:29:23] <Shirley>	AF is "AbuseFilter." You all confuse me.
[18:30:04] <RoanKattouw>	AFT then
[18:30:13] *RoanKattouw 	svn ups the prototype 1.17wmf1 copy
[18:35:50] <RoanKattouw>	Grr
[18:36:03] <RoanKattouw>	Somehow the loader seems to be broken, but I really don't understand how or why
[18:36:37] <RoanKattouw>	It's XHRing the script, but not appending it to the body
[18:36:49] <RoanKattouw>	Even though we do $('body').append('<script ... ')
[18:46:46] <TrevorParscal>	RoanKattouw: so, the math you said was wrong
[18:47:01] <RoanKattouw_>	 Oh of course
[18:47:08] <RoanKattouw_>	Somehow this error wasn't getting thruogh
[18:47:11] <RoanKattouw_>	TypeError: mw.user.id is not a function
[18:47:49] <RoanKattouw_>	How so?
[18:48:15] <TrevorParscal>	hrmph
[18:48:38] <TrevorParscal>	I just fixed the math in r82993
[18:48:58] <RoanKattouw>	Oh you said I was right
[18:49:05] <RoanKattouw>	I parsed your last sentence differently
[18:50:10] <RoanKattouw>	So I will have to merge in the user.id changes after all
[18:50:55] <TrevorParscal>	:) yes, of course you were right
[18:51:54] <RoanKattouw>	Oh BTW
[18:52:13] <RoanKattouw>	When I was in SF last, Neil had some criticism of the method that generates the anon token
[18:52:32] <RoanKattouw>	First was the same as yours, suggestion to use the session key, which I found scary
[18:53:00] <RoanKattouw>	The second was that repeated calls to Math.random() do not actually add any entropy at all
[18:53:24] <Reedy>	lol, randomness
[18:53:38] <Reedy>	RoanKattouw, buy a $10,000 SUN branded random number generator :P
[18:53:48] <RoanKattouw>	haha
[18:53:53] <RoanKattouw>	Not suggesting that
[18:53:57] <RoanKattouw>	But the string could be shorter
[18:54:06] <RoanKattouw>	And we could just derive it from one random value
[18:54:16] <Reedy>	http://upload.wikimedia.org/wikipedia/commons/7/74/Sun-crypto-accelerator-1000.jpg
[18:54:39] <RoanKattouw>	lol
[18:55:04] <Shirley>	http://xkcd.com/221/
[18:55:23] <Reedy>	Shirley, do you suggest repeating calls to that many times for every user? ;)
[18:55:26] <RoanKattouw>	We invoked that one the other day when discussing AFT randomness
[18:55:50] <Reedy>	The not so random random number generator
[18:56:02] <neilk_>	This is not even satire
[18:56:22] <RoanKattouw>	Saying you could use a cutoff for X < page_random < X+n or pageid % X < n where n is a 'guaranteed random by dice roll' number
[18:56:32] <RoanKattouw>	We ended up doing the latter
[18:57:05] <RoanKattouw>	Ah, X is dice roll and n is set according to the percentage you want. Either way.
[18:57:08] <neilk_>	meeting
[18:57:24] <RoanKattouw>	Right, meeting
[18:59:57] <RoanKattouw>	TrevorParscal: Meeting, they're looking for you
[19:00:27] <TrevorParscal>	yes, going now
[19:04:22] <guillom>	jorm, btw, I guess you've noticed, but just to be sure: http://www.mediawiki.org/wiki/Talk:Controversial_content_management_system
[19:12:49] <jorm>	Responded, guillome.
[19:13:37] <guillom>	just saw; thanks
[19:14:11] <RoanKattouw>	Did you guys hear me back there?
[19:14:35] *guillom 	thinks j.orm has a thing for Capitals.
[19:14:41] <guillom>	RoanKattouw, I didn't catch what you said.
[19:14:53] <RoanKattouw>	Oh OK
[19:14:59] <RoanKattouw>	But at least you heard some sound come out? :P
[19:15:03] <guillom>	yes :)
[19:15:08] <RoanKattouw>	I said "Doesn't BZ4 have milestones for this?"
[19:20:18] <RoanKattouw>	TrevorParscal: AFT is now working on prototype, whee!
[19:22:01] <RoanKattouw>	492if(!$.isArray(data.query.articlefeedback)){
[19:22:06] <RoanKattouw>	data.query not defined (API error)
[19:22:08] <RoanKattouw>	Bad Trevor :)
[19:31:18] <guillom>	These meetings always get sidetracked.
[19:48:42] <RoanKattouw>	http://prototype.wikimedia.org/release-en/AricleFeedbackTest
[20:02:28] <RoanKattouw>	TrevorParscal: Did you see https://bugzilla.wikimedia.org/show_bug.cgi?id=23675#c2 ?
[21:28:02] <TrevorParscal>	RoanKattouw: how's it going?
[21:28:29] <RoanKattouw>	Just finished processing invoice and expenses stuff
[21:28:54] <RoanKattouw>	And because the previous one hasn't come through yet, that means I have more than 8 grand in pending invoices now :D
[21:29:51] <guillom>	heh
[21:30:56] <RoanKattouw>	Just converted it to euros and realized it's 1/4 of my taxable income for 2010 in 2 months
[21:36:39] <TrevorParscal>	RoanKattouw: mediaWiki.user.anonymous() was removed?
[21:36:48] <RoanKattouw>	Was it?
[21:36:56] <TrevorParscal>	I'm getting AFT errors about it
[21:37:13] <RoanKattouw>	I merged a ton of things into 1.17wmf1
[21:37:21] <RoanKattouw>	So the error about user.id() would stop
[21:37:36] <TrevorParscal>	yeah, which got it to show up :)
[21:37:57] <RoanKattouw>	You're right, it's not there
[21:38:20] <RoanKattouw>	Running svn blame
[21:38:37] <RoanKattouw>	!r 78544
[21:38:37] --elephant--	http://www.mediawiki.org/wiki/Special:Code/MediaWiki/78544
[21:39:26] <RoanKattouw>	It claims to have been merged, but that's probably the one Platonides messed upo
[21:39:30] <RoanKattouw>	Merging for real now
[21:39:58] <Krinkle>	I'm reading up on MySQL and just wondering if someone could do a quick check if I understand correctly: http://www.youtube.com/watch?feature=channel&hl=en&v=2sBsTruebCQ#t=63m35s This is how Wikimedia has the databases right ? A replicated MySQL cluster for writes and a normal slave replication for reads (best of both worlds)
[21:41:22] <RoanKattouw>	?
[21:41:31] <Krinkle>	the first 60 seconds from 1:03:53 (linked to it)
[21:41:49] <RoanKattouw>	Everything writes to one single master, reads can be from any slave
[21:41:55] <RoanKattouw>	Replication is 1 master --> n slaves
[21:42:21] <RoanKattouw>	The phrase "replicated MySQL cluster for writes, and X for reads" confused me a bit
[21:42:27] <Krinkle>	Sure, it's not exactly the same as the example there but the principle is the same to split reads and writes and have multiple slaves for more complex queries.
[21:42:47] <RoanKattouw>	Reads should go to the slave whenever possible
[21:43:05] <RoanKattouw>	Of course some reads cannot afford to be lagged, those go to the master
[21:43:46] <RoanKattouw>	The diagram in the YouTube video shows a multi-master setup, we don't have those
[21:44:08] <RoanKattouw>	Other than that it's pretty close
[21:44:13] <Krinkle>	ok, nvm. Thanks for the comment.
[21:44:37] <Krinkle>	RoanKattouw: We discusses that revisions before, beware though. I think it had some other issues which was reason not to merge in the past.
[21:44:39] <RoanKattouw>	The only differences are that we have a single-master, not multi-master setup, and that we use sharding (i.e. different wikis on entirely separate DB clusters)
[21:44:41] <Krinkle>	discussed*
[21:44:51] <RoanKattouw>	The "issue" was it wasn't needed
[21:44:58] <RoanKattouw>	Turns out ArticleFeedback does need it
[21:45:05] <Krinkle>	It has dependancies
[21:45:16] <RoanKattouw>	I had to merge those in anyway
[21:45:18] <Krinkle>	athough I usee user.id and user.name are live now. So that's figured.
[21:45:24] <RoanKattouw>	I just merged it, it merged cleanly
[21:45:26] <Krinkle>	Cool