[00:00:15] <RoanKattouw>	Yeah
[00:00:20] <raindrift1>	i could make it block reads, but i don't think that'd work on other DBMSes.
[00:00:21] <RoanKattouw>	You could put the DELETE+INSERT in a transaction
[00:00:33] <RoanKattouw>	That will prevent others from stealing your lock in between
[00:00:40] <RoanKattouw>	You can do it, with LOCK IN SHARE MODE
[00:00:46] <RoanKattouw>	But locking reads are evil
[00:00:53] <raindrift1>	exactly, but that'll definitely be mysql-only.
[00:00:56] <RoanKattouw>	If I deploy locking reads to the cluster, Tim will slap me in person
[00:01:02] <raindrift1>	heh.
[00:01:18] <RoanKattouw>	Which is not as hard as it sounds, given that the reason I'm up at 1am reviewing code is that I'll be on a plane to Australia in 48 hours
[00:01:35] <raindrift1>	i mean, it depends.  i think a locking read would be reasonable here, but since it's possible to do it without, i'll do that instead.
[00:01:51] <raindrift1>	heh
[00:03:17] <maplebed>	brion: is there some magical way in an svn commit to reference a bugzilla bug such that they get html-linked in code review?  (similar to referencing a previous commit?)
[00:03:45] <raindrift1>	oh, right.  that's the thing that blocking gets me.  if i can cause a thread to wait, i can be guaranteed a correct cache.
[00:04:52] <brion>	maplebed: "bug 12345"
[00:05:02] <maplebed>	excellent.  thanks.
[00:05:06] <brion>	same as on bugzilla itself
[00:05:18] <brion>	always nice when syntaxes match ;)
[00:07:30] <petan>	Nemo_bis: you are dev?
[00:07:38] <Nemo_bis>	petan, no
[00:07:53] <petan>	ah, right, anyway I gave you some flags on fi so you can import pages there
[00:18:43] *Joan 	licks Nemo_bis.
[00:19:02] *Nemo_bis 	not sure whether it's a punishment or a compliment
[00:24:15] <Krinkle>	Wow, http://www.wikiquote.org/ looks broken
[00:24:27] <Krinkle>	skins-1.5 @import monobook
[00:25:00] *Krinkle 	 updates
[00:25:10] <Joan>	You can sync it, right?
[00:25:16] <Krinkle>	Yep
[00:25:18] <Joan>	Cool.
[00:25:25] <Krinkle>	Just did it. Using bits.wm.o resourceloader/html5 links now
[00:25:34] <Joan>	Whee. :-)
[00:25:36] <Krinkle>	https://meta.wikimedia.org/w/index.php?title=Www.wikiquote.org_template/temp&action=history
[00:25:47] <Krinkle>	it had 404 errors all over
[00:25:54] <Joan>	https://www.wikiquote.org/
[00:26:00] <Krinkle>	https://meta.wikimedia.org/wiki/Talk:Www.wikiquote.org_template#1.18_update
[00:26:02] <Joan>	All seven visitors were surely disappointed.
[00:26:37] <Nemo_bis>	:(
[00:26:55] *Nemo_bis 	supposed Mxn was still taking care of them
[00:27:12] <Joan>	I think he mostly does wikipeida.org.
[00:27:16] <Joan>	wikipedia.org, even.
[00:27:40] <Nemo_bis>	hmmm
[00:28:17] <RoanKattouw>	Krinkle: Hah, I must've gone through that when doing the prot rel stuff
[00:31:21] <Nemo_bis>	Joan, yes, and Wiktionary
[00:31:31] <Nemo_bis>	stopped to regularly update other projects in like 2008
[00:31:48] <Nemo_bis>	I should update my beliefs about other users' activity more often
[00:33:36] <Joan>	There will be a quiz.
[00:39:12] <Nemo_bis>	Krinkle, did you check other portals as well?
[00:39:17] <Krinkle>	No
[00:39:35] <Krinkle>	Should be similar, I gotta go to bed in a few minutes though
[00:39:41] <Krinkle>	if you have time, I left notes here https://meta.wikimedia.org/wiki/Talk:Www.wikiquote.org_template#1.18_update and diff links
[00:40:04] <Krinkle>	and there's js preview in tab menu
[00:41:23] <Nemo_bis>	yes, I read them
[00:42:01] <Nemo_bis>	but I think I'd need 10-30 times the time you need to do such changes, given that I don't know anything :(
[00:42:35] <Nemo_bis>	looks like at least 404 errors have been fixed by Hoo on 2011-10-19, he forgot only WIkiquote, evil hoo
[01:01:28] <kaldari>	Krinkle: I installed doxygen locally and it looks like @param <type> $var <description> is definitely the right syntax...
[01:01:44] <kaldari>	that syntax gives you:
[01:01:49] <kaldari>	<i>int</i>     <b>$httpRespCode</b>     HTTP response code
[01:02:01] <kaldari>	while the syntax most people have been using gives:
[01:02:07] <kaldari>	<i>$httpRespCode</i>     int HTTP response code
[01:02:34] <Krinkle>	There's probably an option to use the other syntax ?
[01:02:52] <kaldari>	if, so it doesn't look like we have the option set differently
[01:03:10] <kaldari>	I compared against our doxygen output at http://svn.wikimedia.org/doc/ and it's the same
[01:03:15] <Krinkle>	I personally prefer the other one, but that's because of readability (looking for a variable name is easier when they're aligned underneath each other flush against the @param wall
[01:03:59] <kaldari>	I think the problem is the parser has no way to tell where the type ends and the description begins unless you put it before the var name
[01:04:19] <kaldari>	it can identify the var name based on the $, but it has no way to identify types
[01:04:53] <kaldari>	especially since the type can be a custom object or a piped list
[01:05:31] <Krinkle>	Yeah
[01:05:35] <Krinkle>	It could use a colon though
[01:05:43] <Krinkle>	which is used a fair bit in MW as well
[01:05:59] <kaldari>	What does the colon syntax look like?
[01:06:09] <Krinkle>	@param $bar Bar: Description
[01:06:19] <kaldari>	lemme try that...
[01:06:48] <Krinkle>	I haven't seen that outside MW though, could be homebrew
[01:07:08] <Krinkle>	did you find an example of @param type $var Text. at http://svn.wikimedia.org/doc/  ?
[01:07:29] <Krinkle>	Example of @param $var type Text. over here: http://svn.wikimedia.org/doc/classApiBase.html#a6edb411f8465c4f3d50611baaee223e2
[01:07:44] <Krinkle>	<td valign="top"><em>$paramName</em>&nbsp;</td><td>string Parameter name </td>
[01:08:13] <kaldari>	that just gives me <i>$httpRespCode</i>    int: HTTP response code    in my local copy
[01:09:08] <kaldari>	All my code with the other syntax in in extensions :(
[01:09:17] <kaldari>	is in extensions
[01:09:44] <Krinkle>	Anything in core that uses it and is at ./docs/ ? wonder how that is parsed there
[01:09:49] <kaldari>	so I don't know of any examples in http://svn.wikimedia.org/doc/
[01:10:27] <kaldari>	I would be surprised if we're using a customized parsing, but who knows
[01:11:11] <kaldari>	which branch does http://svn.wikimedia.org/doc/ pull the docs from?
[01:11:27] <Krinkle>	trunk
[01:11:30] <Krinkle>	hourly or daily
[01:12:01] <kaldari>	I'll try changing the syntax on 1 function and see what it gives us tomorrow
[01:12:46] <Krinkle>	$ ack 'string $' -Q
[01:13:01] <Krinkle>	There oughta be at least one, if not I think consensus is clear :P
[01:13:12] <Krinkle>	WatchAction.php
[01:13:18] <kaldari>	yes, but consensus can be wrong :)
[01:13:26] <Krinkle>	http://svn.wikimedia.org/doc/classWatchAction.html#a6d8795a4e3ffe4f3d93ab0995a30372a
[01:13:51] <kaldari>	weird, it doesn't bold the param name
[01:14:04] <kaldari>	I guess you're right
[01:14:15] <Krinkle>	<td valign="top"><em>Title</em>&nbsp;</td><td>$title Title object of page to unwatch </td>
[01:14:16] <kaldari>	I wonder if we're using an older version of Doygen
[01:14:22] <Krinkle>	<td valign="top"><em>$paramName</em>&nbsp;</td><td>string Parameter name </td>
[01:14:35] <Krinkle>	I don't think it cares
[01:15:01] <Krinkle>	it puts the 1st word in <em> and the rest in the next
[01:15:18] <Krinkle>	it does magic linking to classnames everywhere regardless of position in description or type
[01:15:24] <kaldari>	Yeah, that's definitely the old Doxygen style
[01:15:53] <kaldari>	so I guess we need to keep doing it that way until we upgrade Doxygen
[01:16:13] <Krinkle>	Put it does space out the variable names in a separate column
[01:16:32] <Krinkle>	http://svn.wikimedia.org/doc/classWatchAction.html#a6d8795a4e3ffe4f3d93ab0995a30372a
[01:16:45] <Krinkle>	There it would make sense if the variable names would be in the first column.
[01:17:16] <kaldari>	the new syntax is for 3-column output: type, var, desc
[01:17:23] <Krinkle>	Yes, that's better
[01:17:37] <kaldari>	our version just does the old 2 col output
[01:18:10] <kaldari>	so "@param $var string: It does this" makes sense
[01:18:23] <Krinkle>	[wikitech-l] Bike Shed 2012-01: @param type $var or @param $var type ?
[01:18:25] <kaldari>	at least for now ;)
[01:19:20] <kaldari>	I'll wait until we upgrade Doxygen before I invite the bikesheding onslaught
[01:19:50] <kaldari>	At least our current syntax is semi-backwards compat
[01:23:28] <kaldari>	or future-compatible I mean
[01:26:02] <Krinkle>	right
[12:55:40] <multichill>	mark: Did you see https://www.ams-ix.net/ams-ix-builds-pop-in-evoswitch-datacenter/ ?
[12:58:06] <mark>	not yet, thanks
[13:07:07] <mark>	but I don't really see a reason to change it soon ;)
[13:51:07] <multichill>	mark: Same here ;-)
[13:51:21] <mark>	we're replacing our foundries by MX80s soon though
[13:51:30] <mark>	i'm glad I left a few RU open below that RX-4 in the new rack
[13:52:46] <multichill>	What did you buy?
[13:53:28] <multichill>	We just bought a bunch of Juniper EX4200 and 4500's. Nice toys :-)
[13:54:33] <multichill>	mark: Going to the ISOC tomorrow btw?
[13:55:15] <mark>	no
[13:55:28] <mark>	and we're buying MX80
[13:59:08] <multichill>	Yeah, I misread it. We also manage some MX80's. We just bough a MX240 ourselves
[14:00:03] <multichill>	Already chose a supplier mark ?
[14:05:27] <mark>	our US one
[17:11:48] <hexmode>	anyone know how to set up upload wizard so it works on deployment?
[17:12:13] <RoanKattouw>	Shouldn't require any special config that I know of
[17:12:50] <RoanKattouw>	Is it broken?
[17:17:55] <hexmode>	RoanKattouw: http://commons.wikimedia.deployment.wmflabs.org/wiki/Special:UploadWizard
[17:23:13] <hexmode>	RoanKattouw: Do you see anyting at that URL besides a spinner?
[17:23:42] <RoanKattouw>	I don't have an account, lemme create one
[17:24:53] <RoanKattouw>	I also see the perpetual spinner
[17:26:01] <RoanKattouw>	And no errors in Firebug, weird
[17:28:03] <petan>	hexmode: LQT works or not?
[17:28:13] <petan>	because this is same behavior as we had
[17:28:16] <petan>	with lqt
[17:28:24] <RoanKattouw>	hexmode: I don't know, looks like UW is just broken. Ask Neil or Ian
[17:28:42] <hexmode>	RoanKattouw: will do
[17:28:43] <petan>	RoanKattouw: I don't think it's wizard
[17:28:51] <petan>	lqt was doing the same
[17:29:07] <hexmode>	petan: lqt url ?
[17:29:11] <petan>	you clicked on new thread and it didn't open, just circle...
[17:29:21] <petan>	hexmode: that was on old wiki before we moved domain
[17:29:26] <RoanKattouw>	Let me try that
[17:29:28] <petan>	new config doesn't have lqt on any wiki afaik
[17:29:33] <petan>	maybe tset
[17:29:35] <petan>	* test
[17:29:35] <hexmode>	k
[17:29:53] <petan>	hexmode: you reported that issue, you know?
[17:30:02] <petan>	lqt didn't work on labs
[17:30:04] <RoanKattouw>	petan: Link to LQT page?
[17:30:25] <petan>	RoanKattouw: that's a problem because we moved wikis to domain and replaced config, so now there is no wiki with lqt
[17:30:33] <petan>	I will enable it on test and will try
[17:30:34] <RoanKattouw>	Oh
[17:30:38] <hexmode>	petan: I don't remember messing with LQT
[17:30:59] <petan>	hexmode: you tried to post some message on labs and told me it doesn't work, it's probably still in irclog
[17:31:06] <petan>	on old
[17:31:15] <petan>	before we moved to labs.wikimedia.deployment
[17:31:16] <hexmode>	petan: ah, yeah
[17:31:18] <petan>	auth
[17:31:22] <petan>	ok
[17:31:24] <petan>	that was same
[17:38:34] <MaxSem>	hey, we have no 1.18wmf1 merge backlog? cool
[17:39:13] <MaxSem>	lemme create some;)
[17:40:03] <petan>	enabled
[17:41:00] <petan>	hexmode: http://test.wikimedia.deployment.wmflabs.org/wiki/Talk:Testghsdh for some reason I can't make it work at all
[17:41:00] <petan>	:D
[17:41:45] <hexmode>	MaxSem: talk to Reedy ... I think he is prepping a release
[17:42:07] <MaxSem>	no, just 1.18wmf1
[17:42:42] <hexmode>	oh
[17:42:45] <hexmode>	yeah
[17:42:58] <MaxSem>	hexmode, already asked but seen no answer - do you still need help with AbuseFilter?
[17:43:00] <Reedy>	Still probably need me to do it ;)
[17:43:32] <hexmode>	MaxSem: if you can do what is needed while Reedy works on the release, maybe?
[17:45:32] <Reedy>	sssslllllooooooooooowwwwwwww
[17:45:38] <MaxSem>	whhhhattt?
[17:46:06] <Reedy>	formey
[17:46:13] <MaxSem>	hexmode, did you say that about backportage or AbudeFilter?
[17:47:57] <hexmode>	MaxSem: if you could get us the AF rulesets (off toolserver?) while reedy does what he needs to do for release, that would help
[17:48:15] <MaxSem>	sure
[17:48:22] <hexmode>	petan: could you import the rulesets or give MaxSem access so he can?
[17:49:42] <MaxSem>	I see no AbuseFilter at http://labs.wikimedia.deployment.wmflabs.org/wiki/Special:AbuseFilter
[17:50:36] <petan>	MaxSem: if you can do that it's great
[17:50:42] <petan>	I was unable to export it from sql
[17:51:44] <petan>	MaxSem: you have access now
[17:52:04] <petan>	MaxSem: sql server is deployment-sql
[17:52:26] <petan>	if you needed I will create sql user for you, or just use service user
[17:52:40] <petan>	and keep in mind that we have no backup of data
[17:53:37] <hexmode>	petan: huh, it looks like MaxSem is right: is AF not installed?
[17:54:40] <hexmode>	petan: I don't see any here, either: http://en.wikipedia.deployment.wmflabs.org/wiki/Special:Version
[17:54:40] <petan>	it's surely installed but probably was disabled by johnd uhart for some reason
[17:54:52] <petan>	just as lqt
[17:55:11] <hexmode>	k, AF is more essential than lqt
[17:55:11] <MaxSem>	due to a crashboombang?
[17:55:21] <petan>	it's enabled
[17:55:28] <petan>	now
[17:55:52] <hexmode>	MaxSem: try again?
[17:56:16] <MaxSem>	see it
[17:56:33] <MaxSem>	I need a +sysop (account name is the same)
[17:56:36] <petan>	sure
[17:56:37] <petan>	wait
[17:57:57] <petan>	can you access it now?
[17:58:10] <petan>	I gave you steward you should be able to access it on all wikis
[17:58:18] <petan>	where it is enabled
[17:58:47] <petan>	if not you can use http://meta.wikimedia.deployment.wmflabs.org/wiki/Special:UserRights to grant  yourself whatever
[17:58:55] <petan>	wherever
[18:01:10] <MaxSem>	hmm, gave myself local sysop, still doesn't work. AF needs extra setup?
[18:01:26] <petan>	probably I will check it, which wiki you try it on
[18:01:50] <MaxSem>	http://labs.wikimedia.deployment.wmflabs.org/wiki/Special:AbuseFilter
[18:02:07] <petan>	hm... ok
[18:02:46] <petan>	I see it
[18:07:16] <petan>	MaxSem: fixed
[18:09:37] <petan>	MaxSem: test it http://en.wikipedia.deployment.wmflabs.org/wiki/Special:AbuseFilter
[18:09:42] <petan>	don't use labs wiki for that
[18:09:47] <petan>	it's not configured for af
[18:09:53] <MaxSem>	sorry, i've a meeting
[18:09:56] <petan>	ok
[18:10:11] <petan>	afk for 40 min
[18:31:18] <causasui>	Hi WikiMedia. Where would I go to float a feature request?
[18:36:41] <RoanKattouw>	For the MediaWiki software?
[18:36:45] <RoanKattouw>	https://bugzilla.wikimedia.org/
[18:39:04] <causasui>	RoanKattouw: Same thing if its en-wiki specific? There is an idea getting traction on Village Pump to add a "report for oversight" button, kinda like the "undo" button
[18:39:20] <RoanKattouw>	"report for oversight" ?!
[18:39:22] <RoanKattouw>	What does that mean
[18:39:44] <guillom>	Some automated way to request to hide an edit, I assume
[18:39:46] <causasui>	Yeah, it's en-wiki specific. On en-wiki we have a group called "oversighters" who can suppress revisions so they are invisible even to admins
[18:39:56] <causasui>	The purpose is to handle illegal or libelous content, etc
[18:40:06] <Ironholds>	RoanKattouw: the RevisionDelete extension :)
[18:40:16] <causasui>	Yeah, kinda like that
[18:40:19] <guillom>	causasui, that seems like an idea for a wiki-specific gadget
[18:40:31] <Ironholds>	causasui: no, technically-speaking revdel and oversight are the same extension
[18:40:42] <Ironholds>	I was just translating "oversight" into something Roan would be more familiar with :)
[18:40:45] <guillom>	Someone should be able to do this in JavaScript
[18:41:09] <guillom>	You should ask WikiProject Gadgets or something like that
[18:41:34] <causasui>	guillom: Well, it's important that it be universal, since the purpose of the proposal is to make reporting for oversight easier without encountering the [[Streisand effect]]
[18:41:41] <causasui>	But I suppose it could start out as a gadget
[18:42:04] <RoanKattouw>	Yeah, this should be easy as a gadget
[18:42:10] <RoanKattouw>	Well, relatively
[18:42:12] <guillom>	causasui, I think not every user needs the tool. It could start as a tool enabled for admins
[18:42:21] <RoanKattouw>	You'd need a page where these requests are collected
[18:42:31] <DarTar>	ping rsterbin
[18:42:36] <guillom>	I think it's usually by e-mail
[18:42:37] <rsterbin>	hey
[18:42:40] <DarTar>	hi there
[18:42:54] <causasui>	guillom: I'm not sure what you mean. Admins can already revdel. The problem is that people who aren't admins don't always report for oversight. See here: https://secure.wikimedia.org/wikipedia/en/wiki/Wikipedia:Village_pump_%28idea_lab%29#Suggestion:_MediaWiki_software_tool_to_flag_revisions_for_oversight.2Frevdel
[18:43:25] <DarTar>	what af_link_id value are we going to use for the new link?
[18:43:50] <rsterbin>	D=4 :)
[18:43:57] <DarTar>	ok thanks
[18:44:00] <guillom>	My point is that if you show this to every possible user, it will add even more complexity to the interface, and may result in an overwhelming amount of requests.
[18:44:01] <rsterbin>	np
[18:44:06] <guillom>	That said, I don't really care.
[18:44:08] <DarTar>	is it documented somewhere other than in the code?
[18:44:24] <causasui>	RoanKattouw: That's also a problem I considered. We don't want to store oversight requests on a page that is visible to non-admins.
[18:44:27] <causasui>	and I don't know if a gadget could accomplish that?
[18:44:44] <rsterbin>	not right now; i've been meaning to update the db section on the tech page, but i haven't gotten around to it
[18:45:02] <guillom>	Basically, you need http://usability.wikimedia.org/wiki/Multimedia:Flows :)
[18:45:35] <RoanKattouw>	causasui: No, you're right, it couldn't. You'd need server-side support for that. Something like flows, yea
[18:45:44] <RoanKattouw>	or queues or whatever
[18:46:13] <causasui>	RoanKattouw: Okay, thanks. Unfortunately I can't develop this due to lack of the necessary technical skill. What pot do you think I should be stirring?
[18:46:48] <DarTar>	rsterbin: ok no problem
[18:48:12] <preilly>	Nikerabbit: you around?
[18:54:02] <AaronSchulz>	maplebed: you'll need to double check the re.match...in case the upload.wikimedia.org bit is part of the string
[18:54:46] <maplebed>	AaronSchulz: the host portion of the url doesn't show up in the path variable.
[18:54:54] <maplebed>	so I'm not sure what you mean.
[18:55:05] <AaronSchulz>	that's what I mean, so good
[18:55:53] *AaronSchulz 	stares at the 'if req.method == 'PUT':' line
[18:56:54] <AaronSchulz>	maplebed: that should just be 'if req.method != 'GET':' right?
[18:57:42] <maplebed>	I don't think so - HEAD should go through translation too.
[18:58:10] <AaronSchulz>	maplebed: do non-auth requests ever HEAD ?
[18:58:15] <maplebed>	but I'm not sure the full spectrum of which HTTP methods need what.
[18:58:34] *AaronSchulz 	checks
[18:58:45] <AaronSchulz>	in any case, certainly nothing other than head/get
[18:58:59] <AaronSchulz>	POST for example updates metadata and is auth only
[18:59:44] <maplebed>	for more details on req.path and req.path_info, you can read /usr/lib/pymodules/python2.6/webob/request.py
[18:59:50] *AaronSchulz 	checks http://www.w3.org/Protocols/rfc2616/rfc2616-sec9.html
[18:59:53] <maplebed>	(to see what parts of the URL are where)
[18:59:58] <AaronSchulz>	yeah, so HEAD/GET is ok
[19:01:12] *AaronSchulz 	changes the code
[19:02:13] <fabriceflorin>	Hey RoanKattouw, rsterbin, DarTar. How's everything coming along for today's AFTv5 release?
[19:02:53] <DarTar>	fabriceflorin: in a conf call right now, but keeping an eye on the channel, ping me if you need something
[19:05:02] <fabriceflorin>	Sounds good, DarTar.
[19:06:22] <rsterbin>	fabriceflorin: i'm looking at some non-release related stuff right now and keeping an eye on the channel -- is there anything you needed before the release?
[19:06:51] <RoanKattouw>	fabriceflorin: I'm about to do the deploy
[19:07:14] <RoanKattouw>	I'm a few minutes late because I was busy carrying a 48-pound suitcase down two flights of stairs :)
[19:07:17] <fabriceflorin>	Hi Reha, nothing on my end for the release. Thanks for being there, in case there are any issues.
[19:07:23] <rsterbin>	no worries
[19:07:42] <fabriceflorin>	Hi Roan, thanks for the update. I take it you found the latest code to work well for you.
[19:08:04] <RoanKattouw>	I haven't actually tested it but it looks good to me yeah
[19:08:31] <fabriceflorin>	Man, 48-pounds is mighty heavy. Is this for your trip to Australia?
[19:08:46] <RoanKattouw>	Yeah
[19:08:54] <AaronSchulz>	maplebed: is it possible to left-anchor the 'hasauth = re.search('/AUTH_[0-9a-fA-F]{32}/', req.path)' regex
[19:08:57] <fabriceflorin>	Good to hear the code looks good to you.
[19:08:58] <RoanKattouw>	I was kind of shocked when I weighed it. Fortunately, the airline's limit is 50
[19:10:03] <RoanKattouw>	oh, hah
[19:10:15] <RoanKattouw>	Looks like you were right and the AFTv5 deployment was scheduled for noon. But whatever, we'll do it now
[19:10:25] <RoanKattouw>	I have two windows reserved, 11am-noon and noon-1pm, and apparently I mixed them up
[19:10:42] <fabriceflorin>	Sounds good to me. Now is fine, if the code is fine.
[19:11:16] <Ironholds>	just tell me when it's live and I'll ping the fuck out of editors :)
[19:11:25] <Ironholds>	sorry, I meant, uh. notify. I'm a professional, honest.
[19:11:59] <maplebed>	AaronSchulz: maybe.  I didn't because it felt like premature optimization.  I don't think we have a performance issue at the moment, and it felt unnecessary.  We would need to look at all the authed urls to verify that it always goes 'v1/AUTH...'.
[19:12:33] <AaronSchulz>	well v1 is hardcoded into cloudfiles
[19:12:43] <AaronSchulz>	I'm more thinking correctness than performance
[19:12:56] <maplebed>	true, but the AUTH part is required for all authenticated swift urls, which are potentially more variable than just cloudfiles.
[19:13:04] <maplebed>	and all swift requests go through this rewrite module.
[19:13:23] <fabriceflorin>	Ha, Oliver. Notify is much nicer. Yes, you ARE a professional ;o)
[19:14:59] <Ironholds>	fabriceflorin: this means I'm not allowed to mix myself a Keyes Special in the next few minutes, right?
[19:15:19] <Krinkle>	Reedy: Can you do me a favor and delete something on mw.org ? https://www.mediawiki.org/wiki/Special:Contestant/1738 looks all messed up and there is no edit or delete button. deleting c24 will fix it
[19:15:45] <Reedy>	err
[19:15:57] <AaronSchulz>	maplebed: you have a '/' check on both sides of AUTH_*, which I guess is ok since uploaded MW files require extensions
[19:16:04] *AaronSchulz 	shrugs
[19:16:06] <Krinkle>	I submitted a fixed one, but page won't behave
[19:16:14] <Krinkle>	looks like the same issue as CR, no tidy
[19:16:18] <Krinkle>	is it that complicated ?
[19:16:23] <fabriceflorin>	Ironholds: No comment. I usually hold off on libations until the main work is done, but that's just me. ;o)
[19:16:35] <Ironholds>	fabriceflorin: a good argument, well made
[19:16:46] <maplebed>	AaronSchulz: basically, I think you're correct that it could be left-anchored, I just haven't done the requisite digging to verify.
[19:17:02] <Ironholds>	a keyes special is a rather unique cocktail consisting of scotch, mixed with scotch, with a side order of scotch to give it flavour. So I should probably avoid it for the time being.
[19:17:15] <Reedy>	Krinkle, done.~ ~~~ doesn't work ;)
[19:17:57] <Krinkle>	Hm.. works for me ?
[19:18:10] <Reedy>	I hope this review was helpful ! ~~~~
[19:18:17] <Krinkle>	Oh, right, *that* doesn't' work
[19:18:17] <fabriceflorin>	I usually pour myself a virgin mary at this time of day (your time). I quit drinking 7 years ago and have never regretted it.
[19:18:37] <Ironholds>	I was going to quit drinking, but I'm not a quitter
[19:18:44] <fabriceflorin>	Hehe.
[19:18:53] <Krinkle>	Reedy: yeah, I guess this is like CR in that it renders live ? So {{REVISONUSER}} and all that fun :P
[19:18:55] <Ironholds>	plus that'd require me to become an alcoholic first if I wanted to do it properly, and I just don't have the time for that kind of commitment, y'know?
[19:19:32] <RoanKattouw>	hmm, crap
[19:19:39] *RoanKattouw 	wonders if the abuse_count and hide_count fields are used
[19:19:42] <rsterbin>	uh oh
[19:19:56] <Ironholds>	now, I don't speak dev. but those don't sound like good noises ;p
[19:19:57] <RoanKattouw>	grr, they are
[19:20:14] <rsterbin>	that's a greg question -- let me ping him
[19:20:25] <RoanKattouw>	I forgot about that schema change
[19:20:30] <RoanKattouw>	They seem to be used, according to grep at least
[19:20:31] <Yoni_OmniTI>	problems?
[19:20:34] <fabriceflorin>	Thanks, Roan. Keep us posted.
[19:20:51] <RoanKattouw>	Hmm, maybe only in the special page? Let's see
[19:20:53] <rsterbin>	greg says they're only used on the special page
[19:21:05] <rsterbin>	whew
[19:21:13] <fabriceflorin>	Yoni, Roan thinks the abuse_count and hide_count fields are used already.
[19:21:36] <Ironholds>	we're saved!
[19:21:49] <RoanKattouw>	yay
[19:21:53] <fabriceflorin>	Well, I'm not sure we were ever lost in the first place.
[19:22:05] <Yoni_OmniTI>	they are only used on the serial page. these belong to comments
[19:22:07] <RoanKattouw>	It does mean though, that if we ever enable that special page, we need a schema change
[19:22:10] <rsterbin>	i was just thinking we should set aside some time to make the update script work
[19:22:13] <Yoni_OmniTI>	special, not serial
[19:22:25] <RoanKattouw>	There's an estimated 33M rows in the table already, so that needs to be a scheduled schema chagne
[19:22:32] <fabriceflorin>	But we have a skeleton for the feedback page that's included in the code, though not visible to users yet. That's probably where these special tags come from.
[19:22:48] <RoanKattouw>	yeah
[19:23:15] <RoanKattouw>	rsterbin: We don't really use the update script on the WMF severs anyways. ALTER TABLE on a 33M-row table requires planning
[19:23:26] <rsterbin>	ok
[19:23:42] <rsterbin>	what do we need to do?
[19:23:54] <RoanKattouw>	Well, 1) figure out when the special page will actually be deployed
[19:23:56] <fabriceflorin>	Sounds good, Roan. Greg will be working on that code for the next couple weeks, towards a Jan. 25 release of the first basic non-public feedback page.
[19:24:11] <RoanKattouw>	Ah, OK
[19:24:13] <RoanKattouw>	Non-public as in?
[19:24:19] <RoanKattouw>	admins only?
[19:25:24] <fabriceflorin>	Non-public as in 'no public links to that feedback page'. (though if people know the URL, they would be able to see that page).
[19:26:18] <fabriceflorin>	It may be possible to limit access based on user level, but we'll make that call as we approach launch, based on how comfortable we are with the first feedback page.
[19:26:41] <RoanKattouw>	OK
[19:26:53] <RoanKattouw>	Well in that case Greg needs to talk to binasher about the schema change
[19:27:26] <fabriceflorin>	Thanks. What's binasher's real world name? WMF employee?
[19:27:37] <RoanKattouw>	Asher Feldman
[19:27:39] <RoanKattouw>	Yeah
[19:27:49] <RoanKattouw>	He's our DBA, essentially
[19:27:57] <RoanKattouw>	I'm not sure if he's officially our DBA but he kind of is in practice
[19:28:14] <fabriceflorin>	Thanks, we'll put Asher and Greg in touch later today. Really appreciate the tip.
[19:29:29] <Ironholds>	so what's the current status?
[19:29:38] <RoanKattouw>	I'm deploying to test now
[19:29:55] *Ironholds 	hides under the desk
[19:31:03] <petan>	causasui: hi
[19:31:11] <causasui>	hi petan
[19:31:13] <petan>	I do enwiki stuff a lot
[19:31:21] <petan>	is there actually a page for reporting that?
[19:31:29] <petan>	I mean oversight related stuff
[19:31:33] <causasui>	Yes, sec
[19:31:43] <petan>	I think this feature can be implemented to twinkle
[19:31:48] <causasui>	petan: https://secure.wikimedia.org/wikipedia/en/wiki/Wikipedia:RFO
[19:31:57] <petan>	ok
[19:32:14] <petan>	don't use secure
[19:32:17] <causasui>	we considered implementing it in twinkle, huggle, and with a gadget. But we had doubts about it, since we weren't sure if such front-end solutions could report the requests for oversight to a private area that only oversighters could veiw
[19:32:23] <causasui>	view*
[19:32:24] <petan>	https://en works
[19:32:55] <causasui>	It's critical that the reports be visible only to the users with the associated rights
[19:33:10] <RoanKattouw>	That's hard, you can't do that in JS alone
[19:33:29] <petan>	causasui: problem is that this information should be private
[19:33:33] <fabriceflorin>	RoanKattouw: Once AFT5 is on test, do we need to create a special article to test it?
[19:33:41] <causasui>	petan: which information?
[19:33:46] <petan>	the tool would need to send an email rather than reporting it to page
[19:34:00] <sumanah>	hi causasui -- I just came in. what are you working on?  I'm WMF's volunteer development coordinator so I like to know what's going on ;)
[19:34:02] <causasui>	RoanKattouw: One suggestion was that twinkle or huggle could load up the "email to oversight" dialogue with something sensible pre-filled
[19:34:10] <RoanKattouw>	fabriceflorin: Not if you did so last time. It was already on there, I'm just updating it
[19:34:22] <petan>	sometimes it's needed to tell sysops to remove such content without doing noise
[19:34:31] <causasui>	Hi sumanah, see https://en.wikimedia.org/wikipedia/en/wiki/Wikipedia:Village_pump_%28idea_lab%29#Suggestion:_MediaWiki_software_tool_to_flag_revisions_for_oversight.2Frevdel
[19:34:33] <RoanKattouw>	Ironholds, fabriceflorin, howief: AFTv5 deployment to test.wp.o done
[19:34:40] <petan>	like posting link to libelous or illegal content isn't prefered
[19:34:52] <Ironholds>	RoanKattouw: awesome :). Nothing melted yet?
[19:34:54] <howief>	awesome
[19:35:04] <RoanKattouw>	petan: The page would have to be something that only oversighters can see, anyway
[19:35:14] <petan>	indeed
[19:35:14] <RoanKattouw>	That was quite clearly established by causasui earlier
[19:35:20] <petan>	that makes it extension worth
[19:35:31] <petan>	ah, didn't notice that
[19:35:32] <RoanKattouw>	Yeah, can't do that in JS alone
[19:35:32] <causasui>	petan, someone had suggested http://usability.wikimedia.org/wiki/Multimedia:Flows if you think it's useful/related
[19:35:36] <sumanah>	https://en.wikipedia.org/wiki/Wikipedia:Village_pump_%28idea_lab%29#Suggestion:_MediaWiki_software_tool_to_flag_revisions_for_oversight.2Frevdel
[19:35:38] <sumanah>	looking
[19:36:00] <petan>	causasui: there is another problem you need to check if someone didn't report it already
[19:36:07] <causasui>	I'm listening
[19:36:15] <petan>	this would require to give limited read access to users
[19:36:23] <petan>	like to check if revision wasn't already reported
[19:36:33] <RoanKattouw>	and to receive responses
[19:36:40] <petan>	yes
[19:36:49] <RoanKattouw>	Like "thanks, I oversighted it" or "you're a troll, go away"
[19:37:16] <petan>	but it sounds interesting we should start a discussion onwiki, causasui can you ask on proposals VP? because this needs a consensus
[19:37:21] <yuvipanda>	sumanah: i'm in touch with potter
[19:37:28] <howief>	how can we view the feedback tool on test?
[19:37:33] <causasui>	petan: I have one going on idea lab, do you think it needs to be moved?
[19:37:35] <petan>	anyway if it has support of community I would like to help with that
[19:37:36] <sumanah>	yuvipanda: got it. Thanks.
[19:37:57] <causasui>	petan: So far, everyone who commented thinks it's a good idea, but it's a small pool. I also shot an email to the audit subcommittee asking them to weigh in
[19:38:13] <causasui>	as if they don't already have enough work :)
[19:38:31] <petan>	causasui: problem of extensions is that if community doesn't like it, it never get deployed and wasting time on programming stuff which is never used is not really what we like to do :)
[19:38:55] <fabriceflorin>	I tried to create this page and URL to test AFT5 on test, but it's not showing the button yet: http://test.wikipedia.org/wiki/Golden-crowned_Sparrow?debug=true&aftv5_show_link=true&aftv5_link=D
[19:38:57] <causasui>	petan: Yeah fair enough. It's encouraging that I haven't yet been able to find anyone who doesn't think it's a good idea
[19:38:59] <petan>	at least if someone like me work on that
[19:39:15] <petan>	that is good
[19:39:44] <causasui>	sumanah, what do you think?
[19:39:52] <petan>	extension should be flexible maybe not only for this purpose
[19:39:54] <rsterbin>	fabriceflorin: firebug has nothing
[19:40:14] <fabriceflorin>	RoanKattouw or rsterbin, is the URL above the correct way to test that AFT5  button is working on test?
[19:40:14] <petan>	so it can be used for more things... it could be useable even on non wmf wikis, maybe
[19:40:16] <rsterbin>	can we check that the category is matching the config
[19:40:17] <rsterbin>	?
[19:40:27] <sumanah>	causasui: I think it sounds like you are writing a spec and getting consensus, and those are good things :-)
[19:40:28] <causasui>	petan: Right, whoever drew up MultiMedia:Flows thought it would have many different uses
[19:40:45] <rsterbin>	fabriceforin: now that i've turned it on by default, you should just be able to go to the page
[19:40:47] <causasui>	This could be just one of tehm
[19:40:59] <fabriceflorin>	I see there is no Article Feedback 5 category on test. Does anyone know how to create a category that would add that article to the category?
[19:41:11] <rsterbin>	RoanKattouw: what's the AFTv5 category on test?
[19:41:29] <sumanah>	fabriceflorin: just marking an article as [[Category:whatever]] should do it, no?
[19:41:30] <rsterbin>	actually, it's probably using the default...
[19:41:33] <rsterbin>	one sec...
[19:41:50] <sumanah>	fabriceflorin: edit the article, place that text somewhere in it, customarily at the end
[19:42:22] <rsterbin>	hm, it's already got the default category set
[19:42:40] <Yoni_OmniTI>	mw.loader.load does not load aftv5
[19:42:41] <rsterbin>	fabriceflorin: it shouldn't matter that the category link is red
[19:42:46] <Yoni_OmniTI>	not configured?
[19:43:05] <rsterbin>	looks like
[19:43:28] <fabriceflorin>	Thanks, Sumanah. I just created this category page: http://test.wikipedia.org/wiki/Category:Article_Feedback_5
[19:43:32] <causasui>	sumanah: Yeah, I think so. I'm just not sure what all the steps are to actually see this through, if other people really agree that it's a good idea. So if you have any pointers on that, I'm listening
[19:44:11] <sumanah>	fabriceflorin: sorry, I don't know
[19:44:15] <fabriceflorin>	But it gives me error messages like 'Wikipedia does not have a category with this exact name.'
[19:44:36] <fabriceflorin>	RoanKattow: What do you recommend we do to test AFT5 on test?
[19:44:45] <RoanKattouw>	rsterbin: whoa, wtf, you're right, it's not even loading the init module
[19:44:52] <RoanKattouw>	Let me look into this
[19:44:54] <Yoni_OmniTI>	localsettings.php?
[19:45:25] <RoanKattouw>	Whoa, wtf
[19:45:27] <RoanKattouw>	Yeah
[19:45:30] <RoanKattouw>	AFTv5 was *not enabled* on test
[19:45:35] <Yoni_OmniTI>	:)
[19:45:36] <RoanKattouw>	Oh! Of course, we used to use enlabs
[19:45:43] <Yoni_OmniTI>	thats right
[19:45:45] <rsterbin>	yeah
[19:45:46] <RoanKattouw>	Can't do that now
[19:45:49] <RoanKattouw>	I'm enabling it on test
[19:46:11] <sumanah>	causasui: hmm -- I think the next step is to write up your spec in a little more detail and circulate it on the Village Pump and to the Wikimedia developers' mailing list, and file an enhancement request in Bugzilla pointing to the spec and the onwiki discussion and any mailing list discussion
[19:46:17] <fabriceflorin>	Yes, enlabs was our last testing environment.
[19:46:56] <fabriceflorin>	So rsterbin, are you suggesting that we can strip out all the debug code from the URL for testing? Remove: ?debug=true&aftv5_show_link=true&aftv5_link=D ??
[19:47:02] <rsterbin>	that's right
[19:47:36] <rsterbin>	i took out the bit of code suppressing the link by default, and D is bucketed at 100%
[19:47:37] <RoanKattouw>	OK, AFTv5 is now enabled on test
[19:47:44] <fabriceflorin>	Got it. Let's try this URL, then: http://test.wikipedia.org/wiki/Golden-crowned_Sparrow
[19:48:10] <rsterbin>	no dice
[19:48:25] <causasui>	sumanah: Well, the discussion is taking place on Idea Lab. Do you think I should try to move it to another one?
[19:48:35] <howief>	it shows up for me now!
[19:48:40] <rsterbin>	cached?
[19:48:48] <RoanKattouw>	Yeah, probably
[19:48:48] <sumanah>	causasui: Ironholds is a better person to ask about that
[19:48:56] <Ironholds>	I am? help!
[19:48:56] <RoanKattouw>	purged
[19:48:58] <rsterbin>	ha, purge=true
[19:49:05] <rsterbin>	got it now
[19:49:07] <rsterbin>	sweet
[19:49:13] <Ironholds>	what's the question?
[19:49:16] <RoanKattouw>	?action=purge purges it from Squid
[19:49:18] <howief>	hmmm
[19:49:24] <rsterbin>	whoops, firefox crashed.
[19:49:30] <howief>	i'm getting some strange behavior on http://test.wikipedia.org/wiki/Golden-crowned_Sparrow
[19:49:32] <fabriceflorin>	I see it too.
[19:49:37] <rsterbin>	ugh, this firebug error is really ticking me off
[19:49:40] <howief>	when i click on the prominent link, i get the overlay (option 3)
[19:49:44] <Yoni_OmniTI>	bucketed for option 3, submitted, no cta
[19:49:57] <Ironholds>	I can see both option 3 *and* D.
[19:50:06] <Ironholds>	sorry, option 2, D. Whichever is the one with praise/criticism/whatnot
[19:50:08] <howief>	i enter feedback, click on Post Feedback, the Post Feedback link gets greyed out and nothing else happens
[19:50:27] <causasui>	Ironholds: I have a proposal on VP:Idea Lab for some changes in MediaWiki software related to oversight. It seems to be popular among people who have reviewed it. Sumanah is suggesting that it needs more circulation on VP, but I'm not sure if the discussion should be moved, or if so, where. The purpose at this point is (a) to get more visibility and (b) through that visibility have a big consensus behind the proposal that I 
[19:50:34] <fabriceflorin>	On Safari, the feedback button gets overridden with a message asking to 'help choose the best picture of the year' at the bottom of the page.
[19:50:36] <rsterbin>	RoanKattouw: firebug's complaining about something in Message
[19:50:45] <RoanKattouw>	Full error msg?
[19:50:53] <Ironholds>	causasui: I would move it to VP(T), and consider sticking something in [[WP:CENT]]
[19:50:59] <rsterbin>	"text is null"
[19:51:08] <RoanKattouw>	How do I reproduce?
[19:51:09] <causasui>	Man, WP:CENT is usually where you go if you want to kill a proposal :P
[19:51:10] <rsterbin>	hang on, let me put it in debug and get a real line number
[19:51:10] <Ironholds>	you could also advertise on wikitech-l if you want some more codemonkeys to take a look at the idea ;)
[19:51:18] <Ironholds>	causasui: I just advertised something there. Thanks :p
[19:51:18] <sumanah>	Ironholds: https://en.wikipedia.org/wiki/Wikipedia:Village_pump_%28idea_lab%29#Suggestion:_MediaWiki_software_tool_to_flag_revisions_for_oversight.2Frevdel
[19:51:28] <rsterbin>	RoanKattouw: i just submitted a comment on option 2
[19:51:33] <howief>	http://imgur.com/HFAYt
[19:51:36] <Ironholds>	sumanah: thanks :)
[19:51:45] <Ironholds>	hmn. you know, we're actually building similar functionality into AFT5.
[19:51:58] <RoanKattouw>	rsterbin: I don't know what option2 is, dumb that down for me please?
[19:51:59] <causasui>	Ironholds: Similar to what? my idea? or is that a second conversation
[19:52:02] <Ironholds>	(well, the end results are similar. Me no grok codemonkeyesque, so the programming may be uuutterly different)
[19:52:03] <Ironholds>	causasui: your idea :)
[19:52:06] <preilly>	AaronSchulz: http://test.wikipedia.org/w/api.php
[19:52:17] <causasui>	Ironholds: Is this something that will launch relatively soon? If so then maybe I can shelve this
[19:52:33] <RoanKattouw>	Ironholds: Similar functionality to what now?
[19:52:39] <rsterbin>	RoanKattouw: option 2 is the form with the suggestion/question/etc tabs
[19:52:40] <thedj>	sumanah: fyi 2
[19:52:43] <fabriceflorin>	Works for me fine on Chrome, but on Firefox I cannot post comments in overlay mode.
[19:52:44] <causasui>	RoanKattouw: My idea about oversight
[19:52:49] <RoanKattouw>	Oh
[19:52:51] <thedj>	regarding Suggestion:_MediaWiki_software_tool_to_flag_revisions_for_oversight.2Frevdel
[19:52:54] <AaronSchulz>	preilly: blame the RFT guys? :)
[19:52:57] <Ironholds>	RoanKattouw: you know the "flag as abuse" thing on the feedback page
[19:52:59] <AaronSchulz>	*AFT
[19:53:01] <RoanKattouw>	No it's only conceptually similar, probably no shared code
[19:53:03] <sumanah>	thedj: sorry, not sure what you mean.
[19:53:06] <Ironholds>	RoanKattouw: what I said :)
[19:53:15] <preilly>	AaronSchulz: can we fix that?
[19:53:18] <Ironholds>	causasui: it is launching soon, but it won't be for article text. so your idea is (distinctly) worth considering
[19:53:19] <thedj>	that's actually a bit like what Krinkle was suggesting in regards with the wikHow ticket based system.
[19:53:32] <thedj>	wikiHow
[19:53:46] <AaronSchulz>	though I wonder why it is doing a DB hit just looking at the api mainpage
[19:53:48] <causasui>	Ironholds: Well, maybe the tack I should be taking is convincing people to roll it in to AFT5 so the same bases get covered (?)
[19:53:55] <thedj>	but then just one ticket queue (proposed rev dels)
[19:54:05] <RoanKattouw>	Whoa, we're hitting a bug in core?
[19:54:32] <rsterbin>	RoanKattouw: jquery.mediawiki module, line 9204
[19:54:38] <RoanKattouw>	Yeah, I got it
[19:54:41] <RoanKattouw>	It's in mw.msg()
[19:54:43] <Ironholds>	causasui: that won't work ;p
[19:54:52] <RoanKattouw>	It dies if you do mw.msg( {object} )
[19:54:54] <Ironholds>	AFT5 is completely unrelated, as a project, from "having a flag-as-abuse option"
[19:54:59] <RoanKattouw>	Which, you know, shouldn't crash
[19:55:06] <rsterbin>	indeed
[19:55:08] <RoanKattouw>	But you also should not feed an object to mw.msg()
[19:55:08] <causasui>	Ironholds: A problem is that article feedback is about articles, but oversight is about revisions
[19:55:20] <Ironholds>	that'd be like trying to roll a design for a fridge into the US army's tank programme because they're both made of metal
[19:55:24] <rsterbin>	that's interesting -- it shouldn't
[19:55:31] <rsterbin>	ok, let me poke
[19:55:33] <RoanKattouw>	rsterbin: jquery.articleFeedbackv5.js line 2551
[19:55:48] <Ironholds>	causasui: they're just completely different things. trust me, you're not going to get the suggestion rolled into AFT5 - although it is a wonderful suggestion :)
[19:56:03] <causasui>	Ironholds: Okay, so your advice is basically the same as sumanah then? Get visibility on VP:T, get consensus behind it, and recruit a dev to implement an extension?
[19:56:07] <Ironholds>	I would advocate setting it up on VP(T) and then poking wikitech-l to see if they have any ideas about how it could be done, technically speaking
[19:56:23] <Ironholds>	causasui: indeed
[19:56:43] <Ironholds>	(I wish I could help more, but my remit is very narrow and includes neither coding skillage nor the ability to order about people who have it)
[19:56:52] <causasui>	All right. I've also shot an email to AUSC asking them to weigh in. Would be nice to hear from some actual oversighters on this. I'll make some noise on VP:T
[19:57:00] <rsterbin>	RoanKattouw: there's probably a db error
[19:57:01] <causasui>	Thanks
[19:57:19] <Ironholds>	causasui: cool beans :). Best of luck!
[19:57:28] <RoanKattouw>	Yes, there is
[19:57:34] <Ironholds>	and I love the idea, so if you need help with on-wiki consensus, poke me and my personal account will be along shortly
[19:58:10] <causasui>	Will do :)
[19:58:15] <RoanKattouw>	Wed Jan 11 19:56:19 UTC 2012    srv193  testwiki        ApiArticleFeedbackv5Utils::getFields    10.0.6.49       1146    Table 'testwiki.aft_article_field' doesn't exist (10.0.6.49)    SELECT  afi_name,afi_id,afi_data_type,afi_bucket_id  FROM `aft_article_field`
[19:58:28] <RoanKattouw>	Did we suddenly start using the _field table?
[19:58:45] <rsterbin>	we've always used that table
[19:59:48] <rsterbin>	RoanKattouw: my fix for the option 3 bug apparently broke the bit that shows you the contents of db errors
[19:59:54] <rsterbin>	i'm going to fix that now
[19:59:59] <rsterbin>	(again)
[20:00:07] <fabriceflorin>	So my current findings are: AFT button seems to work as intended on Chrome 16, Firefox  8 and Safari 5 for Mac OS 10.6. But the feedback forms are not accepting my feedback, no matter which form I use. No Thank you message or CTA is displayed after I click 'Post your feedback'. Anyone else experiencing that issue?
[20:00:12] <sumanah>	thanks Ironholds
[20:00:25] <Ironholds>	sumanah: all I did was go "yeah, what sumanah said" ;p
[20:00:38] <sumanah>	but you can say it with AUTHORITY
[20:00:45] <Ironholds>	I love authority!
[20:00:47] <sumanah>	while I watch a ballerina do impossibly athletic things http://kottke.org/12/01/holy-jesus-look-at-what-this-ballerina-can-do
[20:00:51] <Ironholds>	I love the wooshing sound it makes as it goes above my head
[20:00:57] <thedj>	causasui-away: you should look at the dashboard of wikihow and how it works with collections of tasks that need executing. it would map quite well with your oversight/revdel idea.
[20:01:03] *sumanah 	laughs
[20:01:12] <Ironholds>	okay, now that's just freaky
[20:01:32] <RoanKattouw>	rsterbin: D'oh, testwiki has none of the aft_* tables because it wasn't previously installed there. Holding off on creating them until we've got proper error handling behavior
[20:01:42] <rsterbin>	ok
[20:02:34] <Ironholds>	nobody should be able to stand en pointe like that. Certainly not on someone's arm.
[20:04:47] <rsterbin>	RoanKattouw: checked in
[20:06:34] <sumanah>	Ironholds: btw, tomorrow is the last day to submit a talk to OSCON, one of the big open source conventions, in case you want to submit something about your experience
[20:08:03] <Ironholds>	sumanah: where is it? ;p
[20:08:10] <RoanKattouw>	Ironholds: Portland, Oregon
[20:08:22] <Ironholds>	that is...a wee bit far away for me to travel to
[20:08:23] <Ironholds>	;p
[20:08:29] <Ironholds>	as much as I love conferences!
[20:09:03] <rsterbin>	RoanKattouw: r108665 fixes the error handling
[20:09:08] <RoanKattouw>	Yeah, putting it in
[20:09:20] <rsterbin>	thx
[20:09:23] <Ironholds>	sumanah: thank you for the suggestion anyway, though :)
[20:09:41] <Ironholds>	I'm thinking I might do something more community-centric at wikimania. Y'know, because the two talks I already have just aren't enough
[20:09:56] <sumanah>	Ironholds: sure. :-)  If you were already going to be visiting San Francisco sometime in July, Portland would not be that far.
[20:10:02] <Ironholds>	*nods*
[20:10:11] <Ironholds>	I have no idea if I'll be visiting SF at all, to be honest!
[20:10:12] <RoanKattouw>	Well, Wikimania immediately precedes OSCON this year
[20:10:19] <RoanKattouw>	So you were already gonna be visiting DC in July
[20:10:20] <Ironholds>	and my contract is only to April, so...
[20:10:32] <sumanah>	nod.
[20:10:34] <Ironholds>	RoanKattouw: if I can afford it. It's all very up-in-the-air
[20:10:38] <RoanKattouw>	Right
[20:10:46] <RoanKattouw>	rsterbin: Deployed
[20:11:02] <fabriceflorin>	Thanks, Roan. Am testing now.
[20:11:30] <RoanKattouw>	OK, error reporting works now
[20:11:33] <RoanKattouw>	It says "Form submission error"
[20:11:36] <RoanKattouw>	Now adding the tables
[20:12:08] <rsterbin>	great
[20:12:21] <rsterbin>	(if you have ?debug=true, the sql error shows up in the console)
[20:12:29] <RoanKattouw>	Feedback post went through
[20:12:37] <RoanKattouw>	Yeah, it did
[20:13:06] <fabriceflorin>	I guess I jumped the gun on testing. Roan, please let us know when it's ready for us to test on the front-end. Reha, do I need to empty my cache first?
[20:13:16] <RoanKattouw>	fabriceflorin: It's ready, go test
[20:13:32] <fabriceflorin>	Cool. Worked for me on Safari.
[20:13:36] <DarTar>	fabriceflorin: I note that Google is still running the Nicolas Steno doodle so we should see a crazy amount of posts once the feedback link is deployed
[20:13:54] <DarTar>	do you want me to disable it?
[20:14:08] <DarTar>	it's not going to affect the analysis on the random sample anyway
[20:14:10] <fabriceflorin>	Fine on Firefox and Chrome too. Will test a couple weird use cases now.
[20:14:13] <rsterbin>	whew!
[20:14:21] <fabriceflorin>	Thanks for the heads-up, DarTar.
[20:14:55] <fabriceflorin>	Glad it all worked out, Reha. What was the hang-up? Was it mostly on test? Or was there something wrong in our code?
[20:16:06] <rsterbin>	fabriceflorin: it was the fact that the tables didn't exist on test
[20:16:31] <RoanKattouw>	and that the error-handling code exploded while trying to report that
[20:16:31] <rsterbin>	the js error was because the db error wasn't being correctly handled
[20:16:51] <rsterbin>	yeah
[20:17:07] <fabriceflorin>	That's what I thought. On a different note, what's the proper URL tag for switching buckets? I tried this URL tag, but it doesn't switch forms for me: http://test.wikipedia.org/wiki/Golden-crowned_Sparrow?bucket=2
[20:17:25] <rsterbin>	you need debug=true
[20:18:01] <rsterbin>	if it's not in debug mode, it ignores the url and just uses the bucketing
[20:18:22] <fabriceflorin>	Got it. Works for me with this new URL: http://test.wikipedia.org/wiki/Golden-crowned_Sparrow?debug=true&bucket=2
[20:18:36] <rsterbin>	http://www.mediawiki.org/wiki/Article_feedback/Version_5/Technical_Design#Debug_state_and_query_string_options for more info
[20:18:48] <fabriceflorin>	Thanks. Will test weird use cases now.
[20:18:54] <rsterbin>	awesome :)
[20:19:25] <sumanah>	fabriceflorin: rsterbin: where are you keeping your test cases/test plans?
[20:19:43] <rsterbin>	sumanah: http://www.mediawiki.org/wiki/Article_feedback/Version_5/Testing
[20:19:58] <sumanah>	rsterbin: https://www.mediawiki.org/wiki/Category:MediaWiki_test_plans I'll add that category
[20:20:14] <rsterbin>	neat, i didn't know that existed
[20:20:47] <DarTar>	off to grab a sandwich, bbl
[20:22:10] <sumanah>	rsterbin: ok, just linked to it from https://www.mediawiki.org/wiki/Testing_portal#Test_suites
[20:24:08] <fabriceflorin>	So far, I am not finding any issues on any of the browsers I just tested on my Mac. Could someone please test this on IE8?
[20:24:26] <fabriceflorin>	Thanks, Sumanah, good call.
[20:25:03] <jorm>	here's this: http://www.mediawiki.org/wiki/Echo_(Notifications)
[20:25:11] <jorm>	i need to add a whole "user experience" thing into it.
[20:25:16] <jorm>	and the doc is too technical so far.
[20:25:42] <fabriceflorin>	rsterbin, I did find a minor bug that is not related to this launch: when I click a whole bunch of rating stars in a row in option 3, then click on the trash icon, some of the ratings are not cleared.
[20:26:18] <fabriceflorin>	Other than that, if someone can confirm that his works fine on IE8, I think we are good to go.
[20:26:20] <rsterbin>	bugzilla?  i can look into it after launch
[20:26:34] <rsterbin>	i got mac and ubuntu here
[20:26:39] <fabriceflorin>	Will do. No rush, minor bug.
[20:27:25] <fabriceflorin>	I have asked Erik and Alolita for configuration testing resources for Windows, but haven't heard back from them yet.
[20:27:57] <fabriceflorin>	I don't want to hold off the launch for IE8, but it would help if someone in this chat room could quickly check it at this URL:
[20:28:07] <fabriceflorin>	 http://test.wikipedia.org/wiki/Golden-crowned_Sparrow
[20:30:50] <fabriceflorin>	RoanKattow: From your standpoint, are we ready for deployment of AFT5 on en-wiki? Or are there any tests you think are needed at this time?
[20:31:46] <gregchiasson>	I don't have IE8 handy, but it looks fine in IE9.
[20:32:06] <fabriceflorin>	Thanks, Greg, that's reassuring.
[20:33:09] <fabriceflorin>	Is anyone experiencing any last minute issues with the 'Improve this page' button at the bottom right of the Golden Sparrow page?
[20:33:24] <RoanKattouw>	fabriceflorin: If you say it's good,it's good
[20:33:36] *RoanKattouw 	is on the phone now
[20:33:39] <Ironholds>	seems to be working fine for me. Lets roll this mother out!
[20:33:39] <fabriceflorin>	It's good to go from my standpoint.
[20:33:56] <Ironholds>	or more specifically: lets everyone else roll this mother out, and I'll do the easy work of telling people it has been rolled out and then going to sleep
[20:34:12] <fabriceflorin>	Ironholds, what type of computer do you have? OS? browser version?
[20:34:51] <Ironholds>	firefox 9.0.1, running on windows 7
[20:35:00] <Ironholds>	I can also treat you to firefox on xubuntu, should that be useful ;p
[20:35:27] <fabriceflorin>	Thanks. Do you by any chance have any IE version on your box?
[20:36:00] <fabriceflorin>	IE is truly evil, it makes every developer's life miserable almost every time.
[20:36:24] <Ironholds>	I don't :(
[20:36:37] <Ironholds>	I installed windows 7 and instructed it to obliterate every trace of that foul beastie from my machine
[20:37:10] <rsterbin>	lol, i guess i should call my mom. ;)
[20:38:32] <fabriceflorin>	Good for you. In any case, the fact that it works for Greg on IE9 is comforting. We do not support IE7 in this limited test, so the only one to worry about is IE8. But I think we can go for it even without this test, given that the code has worked well on that platform so far -- and rely on the community to point out any issues.
[20:38:53] <fabriceflorin>	Yes, calling moms could be a solution for this type of testing ;o)
[20:39:19] <fabriceflorin>	In any case, Roan, it's all yours. Deploy when ready.
[20:42:16] <Ironholds>	this is getting very Doctor Strangelove
[20:42:21] <RoanKattouw>	OK, I'm abck
[20:42:38] <RoanKattouw>	rsterbin: Yeah, you're not far off, I was on the phone with my Dad given that I'm traveling to the other side of the world tomorrow
[20:42:47] <rsterbin>	my mom says it works for her on IE8!
[20:43:15] <fabriceflorin>	rsterbin: very cool, thanks so much for calling your mom!
[20:43:19] <RoanKattouw>	lol
[20:43:24] <RoanKattouw>	Two birds, one stone
[20:43:26] <rsterbin>	that is perhaps the silliest thing i have done in the name of testing
[20:43:39] *AaronSchulz 	thinks of animal cruelty
[20:44:12] <fabriceflorin>	Well at least you got to talk to her, which is always a good thing ;o) Please thank her profusely on our behalf.
[20:44:19] <rsterbin>	i did
[20:44:26] <Ironholds>	rsterbin: I was sitting here like "why is she calling her mom? I know deployment can make stuff break, but not dramatically enough to kill the developers or whatever"
[20:44:49] <rsterbin>	i find it entertaining that there are 85 people in here, and nobody has IE8
[20:45:23] <RoanKattouw>	Reedy: Notice: Undefined variable: wmgShortUrlPrefix in /home/wikipedia/common/wmf-config/CommonSettings.php on line 2471
[20:46:22] <Reedy>	It's cause it's set it to null
[20:46:24] <^demon>	rsterbin: I used to have IE4Mac, but Rosetta is no more :(
[20:46:28] <Reedy>	Are we not allowed to default to null? :/
[20:46:31] <rsterbin>	ye gods
[20:49:24] <RoanKattouw>	Deployed a few mins ago
[20:49:27] <RoanKattouw>	Please test on prod
[20:49:37] <RoanKattouw>	Reedy: Apparently not. Use false then
[20:49:58] <Reedy>	Need to check what the extension does...
[20:50:20] <RoanKattouw>	or use false for the wmg and null for the wg if false?
[20:50:46] <Reedy>	lol
[20:51:12] <Ironholds>	so, in english; has it been deployed to en.wp?
[20:51:22] <Reedy>	Ironholds, herp derp
[20:51:40] <RoanKattouw>	Yes
[20:52:27] <fabriceflorin>	Thanks, Roan, that's splendid. Ironholds: let's test a bit on production, before making the announcement.
[20:52:34] <Ironholds>	fabriceflorin: cool beans
[20:52:46] <Ironholds>	just ping me when I need to start telling people; I'll shut up until then :)
[20:56:24] <DarTar>	fabriceflorin: is the ?bucket=N override supposed to work in production with the latest version of AFT?
[20:59:25] <causasui>	Ironholds: Do you think it's worth serious consideration to discuss whether this tool should report to oversight or revdel? Or should I just make out like oversight is the only option we're discussing? I'm on the fence - I don't want it to get shut down because people are split both ways
[20:59:42] <causasui>	and the fact that I am on the fence about it is worrying that that might happen :P
[21:00:15] <Ironholds>	causasui: I would suggest just OS
[21:00:25] <Ironholds>	revdel is not for issues serious enough to really warrant instantaneous reporting
[21:00:31] <causasui>	Ok
[21:00:37] <Ironholds>	plus, there are a comparatively large number of admins, so it's easier to get the attention of one of them
[21:00:49] <JeroenDeDauw>	Ryan_Lane: poke
[21:01:04] <JeroenDeDauw>	Ryan_Lane: I can no longer access my labs instance thing o_O
[21:01:37] <causasui>	Ironholds: I was thinking that the upside was having more admins means more people looking at the queue (faster action) and frivolous reports will be less of an issue with more people clearing them out
[21:01:54] <Ironholds>	causasui: true, but that's irrelevant
[21:01:56] <causasui>	Ironholds: The downside is that more people looking means more people see the stuff that should have gone straight to oversight. Plus, the aforementioned "split the vote" isseu
[21:01:58] <causasui>	issue*
[21:02:04] <JeroenDeDauw>	Ryan_Lane: When I ssh I get "permission denied. public key" and on the wiki it does no longer list any associated stuff w/ me :/
[21:02:18] <Ironholds>	causasui: this is true
[21:02:41] <Ironholds>	however, my point would be that this is all moot; it's about what is ultra-ultra-ultra-important enough to justify instant removal
[21:02:47] <Ironholds>	and that's OSable material, not revdelable material
[21:04:11] <rsterbin>	DarTar: only if you have debug=true
[21:04:13] <causasui>	All right. I wonder though, if suppression is supposed to be only for the most extreme stuff, and we add this button, they'll probably start getting reports for bad things that aren't quite *that* bad
[21:04:32] <causasui>	But that's probably worth it to make sure we're getting everything
[21:04:36] <rsterbin>	RoanKattouw: works great on prod for me
[21:04:52] <Ironholds>	causasui: indeed. and if that does happen and the cons outweigh the pros, we can turn it off :)
[21:04:52] <DarTar>	rsterbin: I thought I was able to rebucket myself in production by appending that parameter, no?
[21:05:04] <RoanKattouw>	DarTar: Apparently you now need debug=true too
[21:05:19] <rsterbin>	yep
[21:05:22] <DarTar>	hmm that's an important change we need to be aware of,
[21:05:33] <DarTar>	we won't be able to send users to a specific version of the widget
[21:05:42] <DarTar>	Ironholds: you aware of that?
[21:05:57] <DarTar>	useful to keep that in mind for the next office hours
[21:06:07] <causasui>	Ironholds: All right, thanks
[21:06:23] <DarTar>	also fabriceflorin ^^
[21:06:32] <fabriceflorin>	RoanKattouw, rsterbin, Ironholds, DarTar: How is the AFT5 button working for you on the main English encyclopedia?
[21:06:44] <rsterbin>	it looks good for me
[21:06:50] <rsterbin>	and works, too :)
[21:07:01] <fabriceflorin>	It's working out great for me for all browsers on articles like these: http://en.wikipedia.org/wiki/Geneva
[21:07:13] <DarTar>	I just noted a discrepancy on Mac OS between Safari and Chrome
[21:07:37] <DarTar>	there is no text decoration on Chrome while I have an underline link on Safari
[21:07:44] <DarTar>	is that a caching issue with my CSS?
[21:08:20] <fabriceflorin>	DarTar, can you clarify what you mean by text decoration? is this on feedback form or CTA or button?
[21:08:26] <DarTar>	the button
[21:08:58] <DarTar>	the link is underlined on Safari, suggesting that a "text-decoration: none" is not respected (I imagine)
[21:09:09] <DarTar>	while it is on Chrome
[21:09:15] <Ryan_Lane>	JeroenDeDauw: let's talk about it in #wikimedia-labs
[21:09:20] <Ryan_Lane>	too much noise in here
[21:09:32] <fabriceflorin>	You guys can test different versions of the form with this URL: http://en.wikipedia.org/wiki/Geneva?debug=true&bucket=1
[21:09:55] <rsterbin>	DarTar: I see it with no underline on Safari
[21:10:00] <rsterbin>	what version are you using?
[21:10:29] <fabriceflorin>	DarTar, can you please send us a screenshot by email? We'll open a Bugzilla bug if we can reproduce on our end.
[21:10:39] <DarTar>	the latest, let me clear my cache first
[21:10:50] <fabriceflorin>	But for now, I don't see that as a showstopper, unless you do.
[21:11:13] <DarTar>	nope, still there after clearing the cache,
[21:11:26] <DarTar>	I'm using Safari 5.1.2
[21:11:41] <rsterbin>	DarTar, me too
[21:11:44] <DarTar>	Mac OS 10.6.8
[21:11:52] <DarTar>	I can send you a screenshot
[21:11:55] <rsterbin>	likewise
[21:13:30] <rsterbin>	DarTar: I've been meaning to ask you -- what are the A, B, and C flags in the top graph of your AFTv5 dashboard?
[21:14:17] <DarTar>	rsterbin: annotations of notable events, just hover over one of those
[21:14:18] <rsterbin>	whoops, never mind, i just managed to catch it on rollover
[21:14:22] <DarTar>	:)
[21:14:50] <DarTar>	let me force a refresh of the graphs to see how we're doing
[21:15:34] <fabriceflorin>	Thanks for the screenshot, DarTar. I am unable to reproduce the problem on my version of Chrome. Is is possible this could be caused by your use of special gadgets?
[21:16:15] <DarTar>	I don't think I've got anything installed on Safari
[21:16:19] <fabriceflorin>	In either case, this seems like a cosmetic bug, not a functional bug, right? Can you still click on the button and get the form? If so, I would try to fix the bug in coming days, but not hold off the announcement by Oliver.
[21:16:35] <DarTar>	yes, it works fine otherwise
[21:17:20] <fabriceflorin>	OK, then I think it's time to wake up Oliver from under his desk and start announcing this release ;o)
[21:18:15] <Ironholds>	I wasn't under my desk!
[21:18:25] <Ironholds>	I was, uh. resting. my. eyes? Yes. that.
[21:18:40] <Ironholds>	I shall start spamming the fuck out of people forthwith :)
[21:18:51] <fabriceflorin>	Ironholds, I was kidding. There was an earlier note about you being under your desk ...
[21:19:24] <fabriceflorin>	Go forth and spread the word ... Please send links to any prominent touchpoints where comment streams might emerge ...
[21:20:17] <DarTar>	only 36 posts submitted since the deployment with af_link_id = 4, I guess the previous version it's still cached for most users
[21:20:41] <fabriceflorin>	RoanKattouw: thanks for all your help with deploying this feature. Have a wonderful trip to Australia and see you in SF in a couple weeks!
[21:20:52] <RoanKattouw>	Thanks!
[21:21:28] <fabriceflorin>	rsterbin: nice work, and thanks so much for being here through this deployment. It's a true pleasure to be working with you -- and your mom!
[21:21:38] <rsterbin>	thanks, you too!
[21:21:40] <Ironholds>	fabriceflorin: will do :)
[21:22:51] <fabriceflorin>	DarTar, would you be so kind to send us all the technical details of the configuration where you are experiencing these problems? I fyou feel so inclined, it might make sense to file this right away on Bugzilla, under aft1.5. Thanks for reporting it!
[21:23:45] <DarTar>	sure, can nobody else reproduce this?
[21:25:01] <DarTar>	ha, interestingly the underline disappears when I log out
[21:25:08] <DarTar>	so it's definitely a gadget issue
[21:26:00] <Ironholds>	fabriceflorin: done! http://en.wikipedia.org/wiki/Wikipedia:Administrators%27_noticeboard#Article_Feedback_Tool_-_additional_test_deployment and http://en.wikipedia.org/wiki/Wikipedia:Village_pump_%28technical%29#Article_Feedback_Tool_-_additional_test_deployment
[21:26:26] <Ironholds>	now, *now* I shall start drinking.
[21:27:08] <Ironholds>	(there are also mailing list posts, but I can't exactly link to those)
[21:33:12] <fabriceflorin>	Thanks, Ironholds, nicely done! Yes, it' time to celebrate. I raise my glass of california fruit juice to you all!
[21:33:36] <Ironholds>	fabriceflorin: what's the difference between california fruit juice and normal fruit juice?
[21:33:51] <Ironholds>	is it more laid back? Is it like "man, I'm getting drunk, but the view is like, GLORIOUS, duuuude"
[21:33:53] <fabriceflorin>	And let's hope that our readers and editors like this new feature.
[21:34:49] <fabriceflorin>	Well, let's see: 4 cups orange juice, 2 cups cranberry juice, 1 banana, 20 grapes, and blackberries, all thrown into a blender. Delicious.
[21:35:19] <Ironholds>	Translation: fruit, fruit, and more fruit. it sounds like a silver bullet, but for students.
[21:35:32] <Ironholds>	"flee! Flee at the sight of my scurvy remedy!"
[21:35:39] <fabriceflorin>	And the view is definitely glorious, but I've been huddled at a computer for four hours and just stepped outside for the first time.
[21:35:58] <Ironholds>	let me again reiterate how much I hate you guys and your reasonable weather
[21:36:35] <howief>	congrats everyone!!!!
[21:37:13] <fabriceflorin>	While this launch stuff was going on, I was also writing a farewell email to 11,000 NewsTrust subscribers to let them know about my new Wikimedia gig, and inform them that we will shut down the NewsTrust.net site in a month. It was tough breaking the news to them, and I have mixed emotions about it all.
[21:37:53] <Ironholds>	can you not get anyone to act as a sort of substitute fabrice? :(
[21:39:25] <fabriceflorin>	Thanks, Howie! I think the launch went well.  We certainly left no stone unturned, and the announcement which Oliver wrote with your help is really clear and to the point. I hope that the community appreciates how hard we work to take their best interests in consideration.
[21:39:44] <Ironholds>	well, mostly it was written by howie with my initial drafting ;p
[21:40:26] <howief>	no i just added a few things here and there
[21:40:33] <howief>	but everything looks great so far
[21:40:49] <howief>	i can't wait to see these comments show up in dario's dashboards
[21:41:36] <fabriceflorin>	Ironholds, I would have loved to find a substitute for NewsTrust, but it also costs us to keep the consumer service going, and my pockets are not deep enough to subsidize that service. But we will continue to offer an educational service, and will donate all our assets to a trusted nonprofit organization to be announced next month. It's all bittersweet for me, but I think it's for the best, so I can focus 150% on
[21:41:36] <fabriceflorin>	Wikipedia.
[21:42:03] <Ironholds>	that is mathematically impossible ;p
[21:43:20] <fabriceflorin>	Well, I'm basing this on the 40-hour workweek which my contract says I'm supposed to work. We all know this will be a 60-hour/week job, because of it's visibility and the huge workload. Fortunately, I like to work ;o)
[21:43:30] <Ironholds>	aww, you're so sweet and naive!
[21:43:38] <Ironholds>	you know, that you think it'll only be 60 hours a week ;p
[21:44:15] <fabriceflorin>	Haha, it better be no more than 60 hours, or my wife will pull the plug. ;o)
[21:44:20] <Ironholds>	that reminds me; howief, my contract informs me I need to get written permission to extend my hours. I've just worked a 10 hour day, and imagine things will be similar in the next couple of days, and I need to do this show-and-tell stuff at the meetup on Sunday. Mind sending that permission through? :)
[21:44:40] <howief>	Ironholds: i'm not sure how that's actually done with o-desk
[21:44:48] <howief>	let me find out
[21:45:00] <Ironholds>	howief: "with difficulty". oDesk actually has me down for 45 a week, but the contract sez 40
[21:45:08] <Ironholds>	and I do not mess with legal. they're terrifying people.
[21:45:34] <fabriceflorin>	HowieF: Yes, I can't wait to see the comments too. But I will probably be really quiet this afternoon, as I huddle down to build the interactive wireframe for the feedback page. I may even shut down my email for a couple hours, so please call me if there is any emergency.
[21:46:11] <fabriceflorin>	Thanks, you guys! I'm going to go get some lunch now.
[21:46:57] <RoanKattouw>	Ironholds: I have that too, it's harmless
[21:47:23] <howief>	fabriceflorin: will do
[21:47:56] <Reedy>	Just get your boss to email HR saying to up your hours on odesk
[21:48:29] <fabriceflorin>	Ironholds: Yes, I had the same issue and Howie fixed it for me. Thanks, Howie -- and bye for now!
[21:50:16] <alolita>	RoanKattouw: are you joining in
[21:50:38] <RoanKattouw>	I'm here
[21:50:47] <alolita>	ok cool :-)
[21:50:49] <RoanKattouw>	It's just that Trevor's talking about stuff I already know
[21:50:55] <alolita>	yup :-)
[22:06:39] <Reedy>	That's 1.18.1 and 1.17.2 out
[22:36:16] <rsterbin>	question for anybody who might know: I need to create a fake article so I can pass text through spam blacklist and abuse filter.  Is there a way to do this by which I can guarantee that I'm not clobbering anything else?
[23:02:51] <RoanKattouw>	rsterbin: Do you really need to create the article though? Don't you just need some wikitext and a Title object?
[23:03:44] <rsterbin>	RoanKattouw: I don't need to put it into the db or anything, but I do need an article object
[23:04:30] <rsterbin>	for abuse filter, i can use that to avoid sending it a title/namespace and risking it coming back with an actual article
[23:04:34] <RoanKattouw>	What's the interface you're trying to use look like?
[23:05:18] <rsterbin>	i don't, alas, have a signature for abuse filter, so i'm guessing based on what it does on edit
[23:05:23] <RoanKattouw>	It sounds like either you're barking up the wrong tree, or AbuseFilter's interface sucks
[23:05:29] <RoanKattouw>	Oh
[23:05:45] <RoanKattouw>	Maybe you should look in the AF code to see if there's a nicely separated function for filtering
[23:05:59] <rsterbin>	that's what i was trying to do
[23:06:02] <RoanKattouw>	I'm fairly confident you'll find it because I know the guy that wrote that code
[23:06:31] <rsterbin>	it takes some variables, but there isn't a list of them anywhere that i can find
[23:06:35] <RoanKattouw>	If it relies on having an Article object as opposed to a Title+content pair, that'd suck
[23:06:52] <rsterbin>	no, it can accept title+content
[23:07:06] <rsterbin>	but it looks up the article, and i want to avoid that
[23:07:19] <RoanKattouw>	Oh, it's fine
[23:07:26] <RoanKattouw>	You probably need to duplicate some code
[23:07:49] <RoanKattouw>	Basically, duplicate most of AbuseFilterHooks::onEditFilterMerged()
[23:08:01] <rsterbin>	yeah, that's what i'm looking at now
[23:08:03] <RoanKattouw>	Except that ACTION and SUMMARY and such would probably be different
[23:08:34] <rsterbin>	i'm having trouble figuring out which keys i'll need to pass in and what they should be
[23:08:45] <RoanKattouw>	Ah, it fetches stuff because there's an oldtext
[23:08:50] <RoanKattouw>	This might be annoying
[23:09:01] <RoanKattouw>	Hmm, nah, it'll be fine
[23:09:06] <rsterbin>	i haven't been able to find what i really want, which is a table of "this key is supposed to have this thing in it"
[23:09:09] <RoanKattouw>	You should look at the ACTION=edit code, but not too closely
[23:09:17] <RoanKattouw>	Instead implement your own ACTION=feedback or whatever
[23:09:24] <RoanKattouw>	It wouldn't have a title or an old_text
[23:09:43] <rsterbin>	i'm not quite sure what that means
[23:09:47] <RoanKattouw>	Hmm OK
[23:09:53] <RoanKattouw>	So you're confused about what the keys in $vars mea?
[23:10:02] <rsterbin>	but then, my brain's a little tired from tracking down a bunch of variables named "var"
[23:10:03] <rsterbin>	yeah
[23:10:17] <RoanKattouw>	The answer is they pretty much mean what you define them to mean, because it's different for each value of $vars['ACTION']
[23:10:32] <RoanKattouw>	For ACTION==edit, there will be old_text, new_text and minor_edit
[23:10:37] <RoanKattouw>	Those things don't exist for e.g. ACTION==delete
[23:10:43] <rsterbin>	ok...
[23:11:05] <RoanKattouw>	If you introduce your own ACTION, you get to make the rules
[23:11:23] <RoanKattouw>	Choose a simple one like delete, move or upload (anything but edit really) and build out from there
[23:11:24] <rsterbin>	right
[23:11:33] <RoanKattouw>	The variables get passed to some sort of scripting language
[23:13:01] <rsterbin>	so, what about parsing the text?
[23:13:17] <RoanKattouw>	What about it?
[23:13:20] <rsterbin>	i know i need title + text in order to use the parser
[23:13:23] <RoanKattouw>	Yeah
[23:13:28] <RoanKattouw>	You can just invent a title there
[23:13:30] <rsterbin>	but i'm not sure what to do about the title bit
[23:13:43] <rsterbin>	i saw that it does a lookup to get existing links
[23:13:48] <RoanKattouw>	It's not really used for anything except having meaningful output for constructs like {{PAGENAME}}
[23:13:52] <RoanKattouw>	Hm, it does?
[23:14:02] <RoanKattouw>	Parser::parse() itself shouldn't be doing that, I don't think
[23:14:03] <rsterbin>	hm, no
[23:14:12] <rsterbin>	that was spam blacklist
[23:14:18] <rsterbin>	whoops ;)
[23:14:52] <rsterbin>	ok, that clears up a lot.
[23:15:11] <rsterbin>	thanks
[23:16:47] <RoanKattouw>	rsterbin: You're dealing with some complex crap there, especially for someone who's a relative outsider to MediaWiki; you have my respect for that :)
[23:17:09] <rsterbin>	heh, thanks
[23:17:23] <rsterbin>	good to know my aching head gets respect ;)
[23:32:20] <preilly>	^demon: ping
[23:32:35] <^demon>	Pong, if it's quick :)
[23:32:39] <^demon>	About to run out for some dinner.
[23:33:03] <preilly>	^demon: what is the correct way to get a  DB object reeference?
[23:33:15] <^demon>	wfGetDB( DB_MASTER )
[23:33:17] <^demon>	Or DB_SLAVE
[23:33:20] <^demon>	:)
[23:33:45] <preilly>	^demon: why is using $wgDBname global bad?
[23:34:39] <^demon>	Aside from the usual mantra of "using globals is bad, use the accessors to them when they're available"
[23:34:56] <preilly>	^demon: okay
[23:35:17] <^demon>	There's also the (possibility) that you could have different $wgDBname's in the course of a request.
[23:35:35] <^demon>	Not practically right now, but I remember Tim & I having a discussion on the issue like 2ish years ago.
[23:53:53] <DarTar>	rsterbin: mail for you
[23:54:04] <rsterbin>	ok, checking...
[23:54:05] <DarTar>	potential problem with clicktracking
[23:55:28] <rsterbin>	let me take a look at that locally -- i'm pretty sure it showed up correctly when i was testing it
[23:56:20] <rsterbin>	huh
[23:56:29] <rsterbin>	it's showing up incorrectly as a bottom impression
[23:56:44] <rsterbin>	i think i can fix that
[23:56:46] <DarTar>	that's what I suspected
[23:56:56] <DarTar>	I see a lot of bottom events that shouldn't be there
[23:57:13] <DarTar>	at least not that many
[23:58:10] <rsterbin>	a hah
[23:58:16] <DarTar>	so was that a misnamed event ID in the code?
[23:58:29] <rsterbin>	it sets the "in dialog" flag *after* loading
[23:58:39] <rsterbin>	i'll swap that round
[23:58:49] <DarTar>	ok
[23:59:31] <DarTar>	can you drop a line in reply to my mail so we can keep track of that? Roan is traveling so I am not sure who could deploy a patch