[00:03:43] <Reedy>	 Ooooh
[00:03:44] <Reedy>	 http://stats.wikimedia.org/wikimedia/squids/SquidReportClients.htm
[00:03:54] <Reedy>	 14 Feb 2012 - 29 Feb 2012
[00:04:11] <Reedy>	 IE6 is 1.53% of all requests
[00:05:26] <Reedy>	 is the magic number 15?
[00:05:31] <Reedy>	 *1%
[00:36:27] <techman224>	 Hello
[04:14:52] <Joan>	 Reedy: It's .01%, I thought.
[04:15:59] <Joan>	 .1$
[04:16:02] <Joan>	 .1%
[10:20:23] <petan|wk>	 @recentchanges-on
[10:20:23] <wm-bot>	 Feed is enabled
[10:22:51] <petan|wk>	 @RC+ mediawiki OAuth
[10:22:52] <wm-bot>	 Inserted new item to feed of changes
[10:34:32] <wm-bot>	 Change on 12mediawiki a page OAuth was modified, changed by 90.183.23.27 link https://www.mediawiki.org/w/index.php?diff=510710 edit summary: /* Relevant bugs */ ++
[10:35:13] <wm-bot>	 Change on 12mediawiki a page OAuth was modified, changed by 90.183.23.27 link https://www.mediawiki.org/w/index.php?diff=510711 edit summary: /* Relevant bugs */ fix
[10:41:52] <petan|wk>	 hm
[10:42:03] <petan|wk>	 @RC+ mediawiki WMF_Projects*
[10:42:04] <wm-bot>	 Inserted new item to feed of changes
[13:50:51] <hexmode>	 10min to 1.19 blocker triage
[13:52:19] <hexmode>	 "1.19 tarball blocker"
[14:01:06] <hexmode>	 anyone here for the triage?
[14:01:58] <JRWR>	 I kinda am, but I'm not sure what the issue even is
[14:02:20] <hexmode>	 JRWR: k, 1min
[14:08:07] <hexmode>	 JRWR: http://etherpad.wikimedia.org/119triage
[14:08:21] <hexmode>	 http://etherpad.wikimedia.org/119triage
[14:08:31] <hexmode>	 anyone else?
[14:11:28] * saper  sits quietly.

[14:13:27] <hexmode>	 saper: yeah, If it is just you two, then I'm not sure how this is gonna go
[14:13:32] <petan|wk>	 hi
[14:13:54] <hexmode>	 petan|wk: hey, stick around, I have questions for you after
[14:14:26] <hexmode>	 fine, 4 people, I say that is a quorum ;)
[14:14:56] <petan|wk>	 ok
[14:15:01] <petan|wk>	 how is https://bugzilla.wikimedia.org/show_bug.cgi?id=34889 possible
[14:15:09] <petan|wk>	 I mean was in normalized ever before
[14:15:17] * hexmode  looks

[14:15:51] <JRWR>	 did someone rewrite it?
[14:16:16] <petan|wk>	 oh, yes it did
[14:16:23] <hexmode>	 petan|wk: micheal m. seems to think it was before... I'm kind of ignorant.
[14:16:24] <petan|wk>	 we should create a 1.18 wiki on labs
[14:16:33] <petan|wk>	 yes it was
[14:16:40] <hexmode>	 petan|wk: exactly what I wanted to talk to you about ;)
[14:16:43] <petan|wk>	 it's hard to check though
[14:16:47] <petan|wk>	 ah
[14:16:54] <petan|wk>	 it's no problem
[14:16:57] <petan|wk>	 werdna: hi
[14:17:01] <petan|wk>	 I think
[14:17:09] <hexmode>	 petan|wk: I used hetdeploy to check out 1.17 and 1.18
[14:17:11] <petan|wk>	 but I need to understand this het - stuff :p
[14:17:34] <petan|wk>	 hexmode: we need to update the db of wikis and recompile it
[14:17:41] <petan|wk>	 no idea how
[14:18:00] <hexmode>	 but yes, getting regressions checked would be
[14:18:03] <hexmode>	 1s
[14:18:04] <petan|wk>	 there is a file to specify which version you use on each wiki in db
[14:18:17] <saper>	 hexmode: what's hetdeploy?
[14:18:19] <hexmode>	 petan|wk: yes, file is updated
[14:18:34] <petan|wk>	 saper: that's some uber thing no one really understand :)
[14:18:42] <hexmode>	 saper: a way to run different versions of mw on one installation
[14:18:56] <hexmode>	 and what petan said
[14:18:58] <hexmode>	 ;)
[14:19:04] <saper>	 hexmode: like, shared parts of LocalSettings.php on one database?
[14:19:09] <petan|wk>	 not really
[14:19:11] <petan|wk>	 that's wiki farm
[14:19:15] <petan|wk>	 we understand that
[14:19:16] <saper>	 or a mega-express-installer
[14:19:35] <petan|wk>	 it's way to configure wiki farm to use different version for certain db
[14:19:55] <petan|wk>	 so if you have a farm of 5 wikis, you can specify a version of mw, you want to use for each of them
[14:20:39] <petan|wk>	 while regular wiki farm, uses one mediawiki folder which contains php code, het deploy enables you to have multiple and assign version to various db's
[14:21:09] <petan|wk>	 in wiki farm, you use same LocalSettings.php for all wikis, only difference is in db's
[14:21:37] <petan|wk>	 saper: make account on labs and I show you how it work
[14:21:46] <saper>	 petan|wk: I have "saper" on labs wiki
[14:21:54] <petan|wk>	 labsconsole.wikimedia.org
[14:22:06] <saper>	 and someone broke my gerrit access, but there is no shell
[14:22:10] <petan|wk>	 ah
[14:22:17] <hexmode>	 1s while I set up a 1.18 wiki -- should have done this sooner :(
[14:22:18] <chrismcmahon>	 petan|wk: one thing I've asked hashar for help with is to make deploying to labs cluster a regular event by way of Jenkins
[14:22:27] <saper>	 petan|wk: I have a running MediaWiki 1.7.3 instance on PostgreSQL on my box ;-)
[14:22:57] <hashar>	 chrismcmahon: hello :-)
[14:23:01] <petan|wk>	 saper: ok, that is good to test heh
[14:23:11] <saper>	 petan|wk: is there are a fast database-populating script, like some articles, some recentchanges, some watchlist, blocks etc.?
[14:23:20] <hashar>	 chrismcmahon: with the git migration, I will not be able to handle that anytime soon :-(
[14:23:24] <petan|wk>	 chrismcmahon: I don't know how jenkins work, but I am willing to help
[14:23:32] <saper>	 petan|wk: what do I need more from labs?
[14:23:45] <petan|wk>	 saper: yes
[14:23:58] <petan|wk>	 there is addwiki script which populate data
[14:24:02] <chrismcmahon>	 petan|wk: it would be useful to build some hetdeploy options into Jenkins builds I think
[14:24:29] <petan|wk>	 I think werdna is likely better person to ask :)
[14:24:37] <petan|wk>	 because he understand het best
[14:24:41] <saper>	 werdna: are you
[14:24:44] <saper>	 good
[14:25:58] <saper>	 hashar: can you check gerrit permissions for test/mediawiki/extensions/examples.git for mere mortals like myself or beau?
[14:26:27] <petan|wk>	 hexmode: link me to 1.18 wiki you made
[14:26:51] <hexmode>	 petan|wk: 1s has turned into 1m
[14:28:54] <hashar>	 saper: any registered user can push a ref there
[14:29:15] <hashar>	 saper: http://imgur.com/bQJgp
[14:29:34] <petan|wk>	 got it
[14:29:44] <petan|wk>	 not really
[14:29:49] <petan|wk>	 http://ee_oldprototypewiki.wikipedia.beta.wmflabs.org/
[14:29:50] <petan|wk>	 ?
[14:29:51] <hashar>	 saper:  refs/for/refs/*  is for gerrit review.   Push and Push Merge Commit  probably mean you can do what ever you want
[14:30:12] <saper>	 hashar: getting permission denied since some time (it used to work)
[14:30:52] <petan|wk>	 hexmode: disable global blocking before you run addwiki
[14:30:57] <petan|wk>	 or database will be fucked
[14:31:04] <hashar>	 saper: would need the exact command being used as well as a paste of the full message :-]
[14:31:09] <petan|wk>	 it's a bug
[14:31:32] <petan|wk>	 hexmode: if u want I can set up that wiki
[14:31:51] <JRWR>	 ive got some full XML dumps if you want to use them
[14:31:59] <JRWR>	 only about 1k pages
[14:32:07] <petan|wk>	 JRWR: for test?
[14:32:11] <JRWR>	 yep
[14:32:18] <JRWR>	 its from my other project
[14:32:18] <petan|wk>	 hm... we could use it
[14:32:31] <JRWR>	 http://pcgamingwiki.com/dumps/
[14:32:36] <petan|wk>	 in fact it would be easier to switch come existing wiki temporary to 1.18
[14:32:43] <petan|wk>	 like simple.wikipedia
[14:32:44] <petan|wk>	 on labs
[14:32:47] <petan|wk>	 it already has db
[14:32:56] <saper>	 hashar: http://tools.wikimedia.pl/~saper/fail/gerrit-fail-01
[14:34:47] <hexmode>	 petan|wk: http://winkyfrown.com/wiki118/index.php/Main_Page
[14:35:06] <hexmode>	 but please do it on labs :)
[14:35:08] <petan|wk>	 ok
[14:35:19] <petan|wk>	 hexmode: what if we switch existing wiki
[14:35:42] <petan|wk>	 or you want to create a new db for it
[14:36:00] <hexmode>	 petan|wk: talk to chrismcmahon
[14:36:10] <petan|wk>	 I mean existing wiki on labs
[14:36:18] <petan|wk>	 like simple.wikipedia.beta
[14:36:21] <petan|wk>	 to 1.18
[14:36:28] <hexmode>	 petan|wk: sounds fine if it works
[14:36:40] <petan|wk>	 let me check
[14:37:38] <saper>	 hexmode: another one, this time "pure" git, no git-review http://tools.wikimedia.pl/~saper/fail/gerrit-fail-02
[14:38:00] <saper>	 s/hexmode/hashar/
[14:38:16] <hashar>	 saper: I have no idea how all those references works
[14:38:17] <hexmode>	 saper: 'bout to say
[14:38:25] <hashar>	 git-review push to refs/for/master/master
[14:38:28] <saper>	 hashar: I figured them out
[14:38:33] <hashar>	 and you are not allowed since  refs/for/refs/*
[14:38:37] <hashar>	 is allowed
[14:38:56] <saper>	 refs/for/master you mean?
[14:39:13] <petan|wk>	 http://en.wikipedia.beta.wmflabs.org/wiki/Special:Version
[14:39:15] <petan|wk>	 problem
[14:39:17] <hashar>	 well it pushes to refs/for/master/master
[14:39:22] <petan|wk>	 Failed opening required 'geshi/geshi.php'
[14:39:23] <hashar>	 but we only allow you in  refs/for/refs/*
[14:39:34] <saper>	 refs/for/refs/* ????
[14:39:36] <petan|wk>	 hexmode: there are some extensions enabled which aren't in 1.18
[14:39:46] <hexmode>	 I'm working on reproducing     http://bugzilla.wikimedia.org/31755 [Regression] Styling in headers breaks sortable tables
[14:39:47] <petan|wk>	 I think we could copy them
[14:39:57] <saper>	 gerrit docs say it's refs/for/<branch>
[14:40:05] <petan|wk>	 or disable
[14:40:35] <saper>	  ! [remote rejected] HEAD -> refs/for/refs/master (branch refs/master not found)
[14:41:44] <hashar>	 saper: try again :)
[14:43:06] <saper>	  ! [remote rejected] HEAD -> refs/for/master (prohibited by Gerrit)
[14:43:21] <petan|wk>	 http://en.wikipedia.beta.wmflabs.org/wiki/Special:Version
[14:43:24] <saper>	 does this thing have ... logs?
[14:43:27] <petan|wk>	 hexmode: ^
[14:43:34] <petan|wk>	 saper: where
[14:43:51] <petan|wk>	 ok, en.wikipedia is 1.18
[14:43:54] <saper>	 petan|wk: gerrit
[14:44:01] <petan|wk>	 saper: yes
[14:44:06] <hexmode>	 ah
[14:44:21] <hexmode>	 \o/
[14:44:28] <saper>	 petan|wk: who can see them? :) maybe there lies secret of my $fail$
[14:44:38] <petan|wk>	 saper: I think you need to ask ops
[14:44:51] <petan|wk>	 saper: Ryan and Roan are good with it, afaik
[14:44:53] <hashar>	 saper: sorry I have no idea what might be wrong :-/
[14:45:12] <petan|wk>	 maybe ^demon knows more, he's git master
[14:46:01] <saper>	 it's too early 7:45 am in SF
[14:46:09] <petan|wk>	 yes
[14:46:27] <hashar>	 ^demon is on vacation, he will be back tuesday
[14:46:44] <hashar>	 I am is back up but I am busy with the git migration :-/
[14:46:49] <hexmode>	 compare: http://winkyfrown.com/wiki118/index.php/Bug_31755
[14:46:54] <hexmode>	 http://winkyfrown.com/wiki/index.php/Bug_31755
[14:47:11] <hexmode>	 ^demon is closer to DC than SF
[14:47:16] <saper>	 hashar: I wanted to check out the possibility to use refs/change/* branches would be cool for colab on small projects
[14:47:18] <hexmode>	 other coast, same time as me
[14:47:39] <saper>	 right, he is in Va. or something
[14:48:47] <saper>	 hashar: isn't having commit access working for gerrit an essential part of git migration? :)
[14:48:49] <hexmode>	 Reedy: http://winkyfrown.com/wiki/index.php/Bug_31755 vs http://winkyfrown.com/wiki118/index.php/Bug_31755
[14:49:35] <hashar>	 saper: yup
[14:49:44] <hashar>	 saper: that is just the test repos which is broken somehow
[14:49:57] <hashar>	 saper: the new repo does not even have a matrix right for now :-]]
[14:51:11] <saper>	 hashar: ah so I am not the only one fixing their git problems with rm -rf <repo> && git clone ... ?
[14:51:18] <petan|wk>	 :D
[14:51:20] <petan|wk>	 saper: no
[14:51:47] <petan|wk>	 in fact I usually have like 3 copies of repository as backup
[14:51:52] <petan|wk>	 so I don't need to wait for clone
[14:52:16] <saper>	 petan|wk: recently learned to nuke my "master" by checking out older commit, nuking master branch with -D and making new master, so other commits disappear from view
[14:52:26] <petan|wk>	 heh
[14:52:42] <saper>	 but they will haunt me there one day I'm sure
[14:54:13] <hashar>	 rm -rf repo should not be needed really
[14:56:30] <petan|wk>	 you are missing skills then, I can make it needed :)
[14:57:15] <saper>	 hashar: there is an error in this http://imgur.com/bQJgp -> refs/for/refs/* should be refs/for/*, can you fix it?
[14:57:27] <hashar>	 saper: nop sorry
[14:57:33] <hashar>	 saper: you will have to wait for someone sosrry
[14:57:52] <hashar>	 I have changed that and it did not fix the issue
[14:57:55] <hashar>	 so that must be something else
[14:58:12] <saper>	 hashar: what's the current value?
[14:58:59] <hashar>	 refs/for/refs/*
[15:01:17] <hexmode>	 got some good regressions going... filling out the etherpad with urls
[15:01:26] <saper>	 hashar: so it's impossible to change that to refs/for/* ?
[15:01:42] <hashar>	 I did and you were still rejected
[15:01:53] <hashar>	 so it is something else and I don't have the time to investigate that any further sorry
[15:02:01] <hashar>	 try with the  test/mediawiki/core repository maybe
[15:02:04] <hexmode>	 https://etherpad.wikimedia.org/119triage
[15:24:40] <hexmode>	 MaxSem: is http://bugzilla.wikimedia.org/35051 fixed?
[15:25:03] <hexmode>	 looks like it
[15:26:17] * MaxSem  svn ups

[15:26:56] <MaxSem>	 hexmode, still broken for me
[15:27:30] <hexmode>	 MaxSem: wfm here: http://winkyfrown.com/wiki/index.php?title=Bug_35051&action=edit&section=1
[15:27:40] <petan|wk>	 hexmode: 34841 is fixed
[15:27:46] <petan|wk>	 it was fixed even before
[15:27:58] <petan|wk>	 http://hub.tm-irc.org/test/w/index.php?title=Main_Page&diff=3085&oldid=3084
[15:28:16] <petan|wk>	 this is latest revision
[15:28:23] <petan|wk>	 if you open any other edit links aren't there
[15:28:30] <petan|wk>	 however 1.19 seems to be broken there
[15:28:36] <MaxSem>	 hexmode, the bug is about summary not being preloaded with /* section title */
[15:28:40] <hexmode>	 petan|wk: svn upping now
[15:28:43] <MaxSem>	 so it's still broken
[15:28:47] <hexmode>	 MaxSem: oh, yeah, sorry
[15:28:49] <hexmode>	 :P
[15:29:08] <MaxSem>	 poor hexmode, too many bug falling on you
[15:29:52] <petan|wk>	 mwaahahaha
[15:29:56] <petan|wk>	 he asked for it
[15:30:59] <petan|wk>	 what I mean is that this bug is already fixed in trunk
[15:31:07] <petan|wk>	 fix isn't merged to 1.19 only
[15:31:25] <hexmode>	 petan|wk: is it tagged 119?
[15:31:42] <petan|wk>	 https://bugzilla.wikimedia.org/show_bug.cgi?id=34841
[15:31:44] <petan|wk>	 this?
[15:32:05] <petan|wk>	 how do I tag it
[15:32:26] <petan|wk>	 oh you mean in svn?
[15:32:30] <petan|wk>	 let me check
[15:32:30] <hexmode>	 petan|wk: which revision, yes
[15:33:29] <petan|wk>	 eeh
[15:33:44] <petan|wk>	 we need to be able to search through commit logs
[15:33:47] <petan|wk>	 :D
[15:34:13] * MaxSem  bisects for

[15:34:15] <MaxSem>	 34841
[15:34:26] <petan|wk>	 bisect?
[15:34:31] <petan|wk>	 what does it mean in simple english :D
[15:34:40] <hexmode>	 petan|wk: part of git sweetness!
[15:34:45] <petan|wk>	 heh
[15:36:12] <petan|wk>	 I think it's faster to fix it than find a fix
[15:36:39] <petan|wk>	 btw the way they fix it in trunk is quite sucky
[15:36:57] <petan|wk>	 I would rather append oldid to edit url
[15:37:12] <petan|wk>	 so that you see edit links but they lead to old revision of page
[15:37:45] <petan|wk>	 question is if it's useful or not :D
[15:38:12] <petan|wk>	 maybe it would let you to revert vandalism and fix typo's from previous non vandal users in one revision XD
[15:38:38] <petan|wk>	 hm...
[15:39:21] <MaxSem>	 okay, you opened a section from an old revision and changed it. what next, into which version of page will you merge it - current or old? in either case, why?
[15:39:37] <petan|wk>	 old, of course
[15:39:47] <petan|wk>	 why? that's a better one :)
[15:39:50] <MaxSem>	 Y????????????
[15:40:06] <petan|wk>	 I don't really know why it should work
[15:40:17] <petan|wk>	 but people don't have to click on that
[15:40:26] <petan|wk>	 for example
[15:40:35] <petan|wk>	 you want to copy a source code of section from old revision
[15:40:38] <hexmode>	 any one want to look at http://bugzilla.wikimedia.org/34889
[15:40:54] <petan|wk>	 MaxSem: ^^
[15:41:09] <petan|wk>	 that's usefull I guess
[15:41:22] * hexmode  doesn't want to disturb MaxSem's bisecting

[15:41:23] <petan|wk>	 how would you copy a source code of section now?
[15:41:33] <petan|wk>	 you would need to open whole page
[15:41:45] <petan|wk>	 and search through code to find where it is
[15:42:02] <petan|wk>	 mediawiki support that function but there is no link to use it
[15:42:25] <petan|wk>	 http://en.wikipedia.org/w/index.php?title=Wikipedia:Village_pump_%28technical%29&action=edit&section=3&oldid=478913156
[15:42:28] <petan|wk>	 look
[15:48:37] <MaxSem>	 fun stuff: close to the end of bisect, I see neither edit link nor the section itself. is it "good" or "bad"? :P
[15:48:49] <petan|wk>	 where
[15:49:11] <petan|wk>	 link I gave you is where the link would point to
[15:53:43] <MaxSem>	 hexmode, tracked down 34841
[15:53:59] <saper>	 !b 34841
[15:53:59] <mw-bot>	 https://bugzilla.wikimedia.org/show_bug.cgi?id=34841
[15:54:15] <hexmode>	 MaxSem: awesome... tagged?
[15:54:21] <MaxSem>	 yup
[15:56:35] <MaxSem>	 investigating 34889
[16:33:38] <MaxSem>	 hexmode, I've fixed 34889
[16:33:57] <hexmode>	 \o/
[16:43:08] <hexmode>	 MaxSem: confirmed fixed
[16:56:40] <sumanah>	 hi Zaran
[17:14:39] <robla>	 Nikerabbit: you around?
[17:14:48] <hexmode>	 RoanKattouw: 20%
[17:14:55] * robla  sees RoanKattouw just got back

[17:14:55] <hexmode>	 where is raindrift??
[17:15:05] <robla>	 hexmode: he's at the product retreat
[17:15:28] <hexmode>	 Just added the still-unclosed 1.19 tarball blockers to http://etherpad.wikimedia.org/20percent
[17:15:35] <hexmode>	 robla: ah
[17:15:44] <robla>	 hexmode: I did some pruning to the list in BZ
[17:15:57] <hexmode>	 we fixed a few of them today
[17:16:13] <hexmode>	 robla: any that had "regression" in the title?
[17:16:33] <robla>	 yeah, at least one, because it was a 1.17 regression
[17:16:55] <sumanah>	 at least, hexmode, we think he is at the product retreat. Slightly worrying since I know he is also giving a talk in Oakland tomorrow afternoon
[17:17:08] <Nikerabbit>	 robla: barely awake
[17:17:14] <sumanah>	 (I should dismiss my nagging feeling that there's a conflict)
[17:17:15] <RoanKattouw>	 Yeah, they know about that
[17:17:20] <sumanah>	 oh good
[17:17:40] <hexmode>	 robla: this one https://bugzilla.wikimedia.org/29277 -- makes sense, I took it off my etherpad just before the triage
[17:18:02] <robla>	 hexmode: I took 31755 off the list too
[17:18:37] <hexmode>	 k
[17:19:16] <robla>	 Nikerabbit: what do you want to focus your 20% time on tomorrow?
[17:19:26] * RoanKattouw  starts doing some code review in the meantime

[17:19:42] * sumanah  thanks RoanKattouw 

[17:20:56] <robla>	 Nikerabbit: one suggestion would be to look at https://bugzilla.wikimedia.org/show_bug.cgi?id=34885 .  Trevor started it, but it needs finishing
[17:21:11] <hexmode>	 is trevor on the retreat?
[17:21:14] <RoanKattouw>	 Yes
[17:21:31] <RoanKattouw>	 Only 1 day, but that's because he's on medical leave on Thursday and Friday
[17:21:53] <RoanKattouw>	 and the entire week after
[17:23:19] <hexmode>	 Nikerabbit: another good one to you to fix or close WONTFIX is https://bugzilla.wikimedia.org/34961
[17:24:00] <hexmode>	 robla: you may have some insight into that one, but it looks like a lot of bikeshedding
[17:24:08] * hexmode  goes to ask siebrand

[17:24:32] <sumanah>	 hexmode: he's gone too
[17:24:35] <Nikerabbit>	 robla: I'm at w3c conference, so no focusing on anything
[17:24:35] <sumanah>	 product retreat
[17:24:41] <Nikerabbit>	 +tomorrow
[17:24:43] <hexmode>	 at least, I would if he were here ;
[17:24:48] <hexmode>	 ;)
[17:24:55] <robla>	 oh....which one?
[17:24:57] <RoanKattouw>	 hexmode: Product retreat!
[17:25:00] <robla>	 Nikerabbit: ^
[17:25:11] <Nikerabbit>	 http://www.multilingualweb.eu/documents/luxembourg-workshop/luxembourg-program
[17:27:24] <Nikerabbit>	 hexmode: re 34961: all the solutions I consider good are rather big to implement, can't do that on normal bug fixing time
[17:27:45] <hexmode>	 Nikerabbit: k, ty
[17:28:33] <hexmode>	 robla: I'm taking it off of 1.19 so we can keep track of it and fix it for 1.20
[17:29:50] <Nikerabbit>	 hexmode: I will keep that bug in mind if I figure out a solution that is easier to implement
[17:30:25] <hexmode>	 :)
[17:30:32] <hexmode>	 down to 4 tarball blockers
[17:30:41] <robla>	 RoanKattouw convinced me that his focusing on code review is a higher priority
[17:31:05] <RoanKattouw>	 I would argue that you convinced me of that just last night, but whatever :)
[17:32:25] <Nikerabbit>	 I've been doing less CR than usual this sprint, but I aim to make up for it next sprint which focuses on bug fixing
[17:33:07] <hexmode>	 hrm... are krinkle and/or Aaron out, too?
[17:33:21] <robla>	 Aaron is here
[17:33:53] <hexmode>	 ah, he wasn't when I checked earlier
[17:34:47] <hexmode>	 AaronSchulz: Any time to follow up on https://bugzilla.wikimedia.org/35047 ??
[17:35:32] <hexmode>	 AaronSchulz: and should that be fixed before we release the tarball?  Seems like it is WMF-centric
[17:39:16] <hexmode>	 AaronSchulz: also, they've updated the bug in the last week since your comment... do you think something could have fixed this?
[17:40:06] <AaronSchulz>	 possibly
[18:12:04] <AaronSchulz>	 hexmode: is http://www.winkyfrown.com/coverage supposed to be up to date?
[18:12:37] <hexmode>	 AaronSchulz: hrm... don't think so... let me check
[18:13:22] <hexmode>	 AaronSchulz: no, that should be on jenkins, isn't it?
[18:13:37] * AaronSchulz  is looking

[18:15:45] <hexmode>	 AaronSchulz: def should be, but I can't find it.  Creating a bz request now.
[18:15:49] * AaronSchulz  gives up

[18:19:27] <hexmode>	 AaronSchulz: already there https://bugzilla.wikimedia.org/31518
[20:40:27] <hashar>	 Ryan_Lane: I would need a review of a simple script to list mediawiki extensions we are hosting
[20:40:33] <hashar>	 that is change https://gerrit.wikimedia.org/r/#change,3149
[20:42:49] <hashar>	 or let me know if someone else can review it :-]
[20:44:04] <RoanKattouw>	 Hmm
[20:44:08] <RoanKattouw>	 I would beef that PHP script up a bit
[20:44:24] <Ryan_Lane>	 why is this on the gerrit server?
[20:44:32] <RoanKattouw>	 Trim the .git , HTML-escape the extension names
[20:44:37] <RoanKattouw>	 Maybe format as an HTML list
[20:45:02] <Ryan_Lane>	 ah. I see why
[20:45:18] <Ryan_Lane>	 this means the gerrit server needs to have php installed
[20:45:20] <hashar>	 the git repos are on gerrit and ls-project  is not publicly available
[20:45:39] <hashar>	 ohhhh no php there :/
[20:45:51] <hashar>	 we could do it in perl :-]
[20:45:53] <RoanKattouw>	 It could be some other CGI-like thing, but then you'd have to install that
[20:46:08] <Ryan_Lane>	 why not just have a cron that runs on the system that generates a static file?
[20:46:24] <hashar>	 RoanKattouw: I just need an up to date list. Any processing will be done by another script.
[20:46:42] <RoanKattouw>	 No it won't, not if you're exposing this to the web
[20:46:42] <Ryan_Lane>	 I think regenerating a file every minute is up-to-date enough
[20:46:47] <Ryan_Lane>	 and then it's a simple bash script
[20:46:53] <Ryan_Lane>	 hell, it isn't even a script
[20:46:54] <RoanKattouw>	 HTML-escaping and setting a Content-Type is the very very least you need to do
[20:47:01] <RoanKattouw>	 But yeah updating a static file works too
[20:47:39] <Ryan_Lane>	 ls −1 > /var/www/mediawiki-extensions.txt
[20:47:51] <hashar>	 RoanKattouw: why should we need to HTML escape / content type that list ?   Are you afraid that someone create a repo whose name can be interpreted by internet explorer as a .exe ?
[20:47:55] <hashar>	 RoanKattouw: I am confused :)
[20:48:03] <hashar>	 Ryan_Lane: yeah that would do it
[20:48:05] <RoanKattouw>	 Not a .exe but HTML/JS
[20:48:33] <hashar>	 RoanKattouw: that would need someone to create a repository named <script>some js there</script>
[20:48:37] <RoanKattouw>	 Yes
[20:48:40] <hashar>	 which is veryyy unlikely to me
[20:48:49] <RoanKattouw>	 Well it shouldn't be allowed
[20:48:54] <RoanKattouw>	 of course
[20:49:00] <RoanKattouw>	 But you should defend against that close to the output
[20:49:18] <RoanKattouw>	 I know that IE will still interpret Content-Type: text/plain as HTML if it wants to
[20:49:26] <hashar>	 Ryan_Lane: I am abandoning the change and making it a cronjob. will be wayyy easier
[20:49:28] <RoanKattouw>	 not sure if that still happens if it's also got .txt as its extension
[20:50:01] <Ryan_Lane>	 ok
[20:50:08] <Ryan_Lane>	 though...
[20:50:15] <Ryan_Lane>	 you don't *have* to abandon it
[20:50:26] <Ryan_Lane>	 you can git reset —hard origin/production
[20:50:29] <Ryan_Lane>	 then make your changes
[20:50:35] <Ryan_Lane>	 then commit, and use the same change-id
[20:50:48] <Ryan_Lane>	 then it'll push it into this change
[20:50:56] <hashar>	 yeah amending
[20:52:45] <RoanKattouw>	 Fuck you IE
[20:53:04] <RoanKattouw>	 Even if you create foo.txt and serve it with Content-Type: text/plain , <script> tags are still run
[20:53:09] <hashar>	 ask brion about IE :-]
[20:53:38] <hashar>	 RoanKattouw: you know there is only a very few people which are able to create git repos in gerrit ?
[20:53:48] <RoanKattouw>	 Yes
[20:53:58] <hashar>	 RoanKattouw: and that gerrit probably only accept [a-Z0-9] as valid chars for a repository name :)
[20:54:04] <RoanKattouw>	 "probably"?
[20:54:07] <RoanKattouw>	 I'd like to know for sure
[20:54:28] <RoanKattouw>	 I know I'm being paranoid here, but I'm also being serious
[20:54:38] <hashar>	 what I mean is that who ever is able to create repository can already do much nastier stuff
[20:55:02] <RoanKattouw>	 Ryan_Lane: What are the plans for doing repo creation in The Future?
[20:55:24] <Ryan_Lane>	 RoanKattouw: we should add support to mediawiki for it
[20:55:34] <Ryan_Lane>	 then we can do it in a controlled manner
[20:55:44] <RoanKattouw>	 OK
[20:55:45] <Ryan_Lane>	 and could let people create their own repos
[20:56:20] <Ryan_Lane>	 it shouldn't be too terrible to add
[20:57:03] <Ryan_Lane>	 I really need to finish my latest OpenStackManager changes before I add something new to labsconsole, though
[21:13:55] <hashar>	 Ryan_Lane: RoanKattouw : a much simpler change https://gerrit.wikimedia.org/r/#change,3149  :)
[21:14:22] <hashar>	 ahh I forgot to redirect the ouput :/
[21:14:24] <Ryan_Lane>	 hahaha
[21:14:27] <Ryan_Lane>	 I was just about to say that
[21:14:48] <hashar>	 I am coding while watching TV and speaking to my wife about our summer vacations ...
[21:14:52] <RoanKattouw>	 Yeah this will mail the list to root@wm.o 96 times a day
[21:14:52] <hashar>	 does not really help
[21:15:11] <RoanKattouw>	 There's enough cron spam going to that address as it is
[21:15:20] <hashar>	 user gerrit2 so ?
[21:16:14] <hashar>	 if there is no output on std err why does it send a mail to root?
[21:16:48] <RoanKattouw>	 Hmm maybe that is stderr only, I don't know
[21:16:57] <RoanKattouw>	 I thought stdout also went there
[21:17:10] <Ryan_Lane>	 I think it's stderr
[21:17:18] <RoanKattouw>	 Oh and it runs as gerrit2 anyways I guess
[21:17:19] <Ryan_Lane>	 it could also be stdout, though
[21:17:29] <Ryan_Lane>	 I though it said root
[21:17:33] <hashar>	 it says root
[21:17:40] <Ryan_Lane>	 it should likely be root
[21:17:54] <hashar>	 I have set it with   user => root
[21:18:04] <RoanKattouw>	 OK
[21:18:08] <Ryan_Lane>	 otherwise we have to add a directory that's owned by gerrit2
[21:19:20] <hashar>	 patchset4 https://gerrit.wikimedia.org/r/#change,3149
[21:21:28] <hashar>	 thanks Ryan :-))))
[21:21:47] <Ryan_Lane>	 yw
[21:22:23] <Ryan_Lane>	 force running it on manganese
[21:22:40] <hashar>	 have you ever used "git archive" ?
[21:22:49] <hashar>	 I tried it this morning, could not make it to work with gerrit
[21:23:08] <hashar>	 I need to fetch files in each of those repositories, and I am doing it with gitweb and the raw format  :-]
[21:23:18] <hashar>	 (to avoid cloning all repositories)
[21:23:47] <Ryan_Lane>	 never used it
[21:23:57] <hashar>	 was just wondering
[21:24:07] <hashar>	 my script already use gitweb to fetch the files anyway
[21:24:18] <Ryan_Lane>	 umm
[21:24:22] <hashar>	 it is going to be ugly :-]
[21:24:28] <Ryan_Lane>	 is this *really* the output you wanted?
[21:25:08] <hashar>	 well to me raw means unmodified in anyway?
[21:25:31] <hashar>	 then I write the fetched content straight in a file
[21:25:43] <Ryan_Lane>	 hm. it's also not showing up
[21:26:15] * hashar  checks docroot

[21:26:45] <Ryan_Lane>	  File does not exist: /htdocs
[21:26:46] <Ryan_Lane>	 weird
[21:27:28] <hashar>	 the :80 virtual host has /var/www
[21:27:37] <hashar>	 the :443 has no DocumentRoot set
[21:27:44] <hashar>	 files/apache/sites/gerrit.wikimedia.org
[21:33:06] <hashar>	 Ryan_Lane: I need to get to bed.  I will ask Ariel to have a look at it tommorow
[21:33:21] <hashar>	 Ryan_Lane: thanks for the cron  / shell one-liner  suggestions :-))))))))
[21:36:50] <Ryan_Lane>	 hashar: yw
[21:37:07] <Ryan_Lane>	 ah, missing docroot would be the problem
[21:51:50] <Ryan_Lane>	 hashar: I fixed your one liner, btw, and fixed apache: https://gerrit.wikimedia.org/mediawiki-extensions.txt
[21:51:53] <hashar>	 Ryan_Lane: thanks for fixing my crappy code :)
[21:52:30] <Ryan_Lane>	 heh
[21:52:35] <Ryan_Lane>	 yw
[21:53:03] <hashar>	 oh and using sed
[21:53:23] <hashar>	 I thought it would get rejected for being too long :-]
[21:53:40] <hashar>	 that list is great. Thanks again Ryan, you are going to save the i18n team
[21:53:45] <hashar>	 see you tomorrow!
[21:53:51] <Ryan_Lane>	 see ya
[21:55:05] <hashar>	 that was bug 35223
[21:55:17] <hashar>	 still have ~ 7000 bugs  to fix
[21:55:20] <hashar>	 bye bye
[22:04:54] <OrenDsk>	 sumanah: re
[22:06:07] <OrenDsk>	 are you arranging berlin ?
[22:06:24] <OrenDsk>	 sumanah: I mean the hackathon
[22:06:46] <sumanah>	 hi OrenDsk! I am.
[22:07:23] <OrenDsk>	 not much info about it online
[22:07:44] <sumanah>	 OrenDsk: not yet, no
[22:08:10] <sumanah>	 OrenDsk: there will be more on Friday.
[22:08:14] <OrenDsk>	 I am thinking to come
[22:08:19] <sumanah>	 OrenDsk: but you should keep that space free on your calendar
[22:08:19] <sumanah>	 good!
[22:08:44] <OrenDsk>	 it would be great if there was some help with accommodations
[22:09:24] <sumanah>	 OrenDsk: Yes!
[22:13:01] <sumanah>	 I need to go, OrenDsk, but when we open registration, there will be an option to say "I want help with accommodation"
[22:13:09] <OrenDsk>	 ok
[22:13:55] <OrenDsk>	 I asked  for a pupet fpr dummies session
[22:14:10] <OrenDsk>	 I'll be around most of 2morrow
[22:14:16] <OrenDsk>	 long weekend
[22:14:20] <JRWR>	 I use nginx for static cache, and it seems that mediawiki is setting session cookies on first page load, wanted to know if there was a option about this anywhere
[22:53:48] <ashish_d>	 RoanKattouw:
[22:53:54] <RoanKattouw>	 ashish_d: Sorry about that, #mediawiki is kinda noisy right now
[22:53:58] <RoanKattouw>	 So yeah, parent hash thing
[22:54:44] <RoanKattouw>	 I think we had a linear version numbering scheme in mind
[22:55:22] <RoanKattouw>	 Because, at least the way I imagine it, there's a central server that maintains a linear history of the document (i.e. no branching/merging) which is authoritative
[22:56:31] <RoanKattouw>	 So, more generally, Trevor and I talked about collaborative editing a bit this week, and I'm not sure if he relayed this to you yet
[22:56:59] <RoanKattouw>	 Basically my idea, which he agreed on, was that it would be really cool if someone implemented collaborative editing but stopping just short of the actual conflict resolution / OT part
[22:57:47] <RoanKattouw>	 So you'd have a Node.js server that receives changes from a client and broadcasts them to other clients, and that would allow people to be a spectator on other people's editing sessions (read-only)
[22:58:13] <RoanKattouw>	 That's quite a bit of work already and it involves all the foundations you need for collaborative editing except conflict resolution/OT
[22:58:38] <RoanKattouw>	 And it results in a cool feature: you could watch someone else edit in real time, and you could even have sessions where you "pass the torch" and let someone else drive for a while
[22:59:23] <RoanKattouw>	 Basically you'd have collaborative sessions where only one person can be the "editor" at any time, and the others are just watching. This is almost as useful as real collaborative editing, so it's a cool thing to put out there, and I think it makes sense to split up the work that way
[22:59:25] <RoanKattouw>	 Thoughts?
[23:01:19] <ashish_d>	 I get that but its kinda hard for me to take it as collaborative editing, like you said other can just watch the changes being made live
[23:01:31] <ashish_d>	 ofcourse, that could be taken forward once this is done
[23:01:41] <RoanKattouw>	 Yeah, it's not "real" collaborative editing
[23:01:47] <RoanKattouw>	 It's really just an intermediate stage
[23:02:01] <RoanKattouw>	 But I think that of all the intermediate stages possible, it's the most useful one
[23:02:06] <RoanKattouw>	 Think of it as phase 1 and phase 2
[23:02:52] <ashish_d>	 so, do you think phase 1 would suffice for a GSoC project? coz as far i can visualize that cuts "alot" of work
[23:03:22] <RoanKattouw>	 It is a hell of a lot of work
[23:03:40] <RoanKattouw>	 Phase 1 by itself is probably more than enough for a GSoC project
[23:03:47] <RoanKattouw>	 You could add phase 2 as an "if time permits" thing
[23:04:22] <RoanKattouw>	 And, you know, if you want to continue working on it post-GSoC (or start working on it early, pre-GSoC) you'd be more than welcome to
[23:04:59] <RoanKattouw>	 MediaWiki as a whole has lots of volunteer developers (or developers not associated with the Foundation), so that's totally normal
[23:05:29] <ashish_d>	 So, my GSoC project would actually say, Implement the initial collaboration server which serves the document to new clients and the broadcastes changes made to it real-time
[23:05:41] <RoanKattouw>	 Yeah, that sounds good
[23:06:27] <RoanKattouw>	 And then if for some reason you have more time, you could do the phase2 thing. I think that would be good to mention in your proposal as well (either as in "this is what this leads to eventually" or "this is what I will do if I have more time")
[23:07:33] <ashish_d>	 sounds good to me, have had a few OT nightmares already :P
[23:08:03] <RoanKattouw>	 hehe
[23:08:46] <ashish_d>	 do you know I also started working on a raw collaboration server?
[23:09:12] <RoanKattouw>	 Well I mean this project it so huge, you'll very likely need to cut it up /somehow/ . And having the infrastructure in place without OT is probably more useful than half-working OT that you never had the time to finish
[23:09:17] <RoanKattouw>	 No?
[23:09:19] <RoanKattouw>	 Tell me more :)
[23:10:53] <ashish_d>	 ok, so I started with a node.js server using socket.io to maintain connections with the clients, built a communication api, to you know receive and broadcaste the transactions
[23:11:17] <ashish_d>	 then  a client adapter which hooked into the surface listening to the new transactions
[23:11:52] <ashish_d>	 and eventually, some realtime collaboration worked by naively communicating the transactions among the connected clients
[23:12:09] <ashish_d>	 https://github.com/dash1291/VE-collaboration this is where I've been experimenting
[23:13:11] <ashish_d>	 and lately, I've been trying to integrate VE modules on to the server, so that the server is able to serve the document by itself and also apply the incoming transactions to it(so far the document came from the VE sandbox)
[23:14:12] <ashish_d>	 pretty much done it
[23:14:51] <RoanKattouw>	 Nice
[23:15:29] <RoanKattouw>	 Hmm, in client.js I think you may need to apply the transaction to the data model instead of to the ES component
[23:15:34] <RoanKattouw>	 (BTW, we are moving from ES to CE)
[23:15:48] <RoanKattouw>	 But Trevor has a more high-level view of this project, he'll know better than I
[23:15:51] <RoanKattouw>	 I'm just the data model guy
[23:16:37] <ashish_d>	 yeah, when I got stuck with this polling thing of transactions, he told me CE handles that pretty well
[23:16:46] <RoanKattouw>	 Yeah
[23:17:12] <RoanKattouw>	 Ideally you would just apply stuff to the data model, and CE or whatever display surface is active would listen for events coming from the data model and render things
[23:17:22] <RoanKattouw>	 I think that's how it's supposed to work
[23:17:46] <RoanKattouw>	 (brb)
[23:22:31] <RoanKattouw>	 (back)
[23:22:56] <RoanKattouw>	 ashish_d: So maybe you've told Trevor already, but where are you located, geographically?
[23:23:19] <ashish_d>	 New Delhi, India
[23:23:22] <RoanKattouw>	 Aha
[23:24:28] <JRWR>	 RAWR, I hate debugging MW's cache/session system
[23:24:48] <RoanKattouw>	 There will be a hackathon in Berlin in early June that some of our team will attend, and we'll be at Wikimania in Washington, DC in July. We also regularly have events in India, I forget what&when the next one is
[23:25:18] <ashish_d>	 yeah, I attended Pune Hackathon, the previous month
[23:29:17] <RoanKattouw>	 !events
[23:29:17] <mw-bot>	 We run face-to-face events for MediaWiki developers and people who want to learn: https://www.mediawiki.org/wiki/MediaWiki_developer_meetings We also have online meetings in IRC to triage bugs: https://www.mediawiki.org/wiki/Bug_management/Triage
[23:31:22] <ashish_d>	 RoanKattouw: So, do you think I should start writing my application, revamping the proposal keeping in mind our discussion?
[23:31:36] <RoanKattouw>	 Yes, please do
[23:31:45] <RoanKattouw>	 I was just talking to Sumana, she said Wikimedia India puts on lots of event
[23:31:47] <RoanKattouw>	 s
[23:31:59] <RoanKattouw>	 Like, smaller local events
[23:32:31] <ashish_d>	 yeah, indeed, I heard another hackathon near the fall in Bangalore
[23:32:45] <RoanKattouw>	 Large events that WMF flies in SF people for happen 2 or 3 times a year, but apparently smaller local events (like one just for New Delhi people) apparently happen frequently
[23:33:05] <RoanKattouw>	 Yeah I hear rumors that there will probably be another large event in India in the fall
[23:33:16] <RoanKattouw>	 and it's actually sort of likely that I would go there
[23:33:52] <RoanKattouw>	 Just sayin', you wouldn't find any visual editor team members at, say, a local New Delhi meetup, but talking to other Wikimedians is fun
[23:33:54] <ashish_d>	 cool! really hope to see you there,  I have even written scholarship application for Wikimania blindly hoping for it :P
[23:34:00] <RoanKattouw>	 Ah, sweet!
[23:34:18] <RoanKattouw>	 Hopefully you'll get accepted and we'll meet there :)
[23:34:30] <ashish_d>	 that would be awesome! thanks
[23:35:29] <RoanKattouw>	 Do you have a link to the current version of your proposal?
[23:36:34] <ashish_d>	 urm, I actually just shot my proposal on wikitech-l, among other stuff I have been keeping notes here http://www.mediawiki.org/wiki/User:Dash1291/Thoughts
[23:36:57] <ashish_d>	 apart from that, I haven't actually drafted my proposal somewhere else
[23:37:29] <RoanKattouw>	 Ah yes, wikitech-l post dated Feb 19
[23:37:36] <ashish_d>	 yep!
[23:37:47] <RoanKattouw>	 That was the day I moved halfway across the world, so I wasn't paying much attention at the time
[23:38:13] <ashish_d>	 heard some of it in the channel ;)
[23:39:08] <RoanKattouw>	 I just read through the thread, I now remember half-reading this
[23:39:26] <RoanKattouw>	 (FYI, a permalink to your post is http://lists.wikimedia.org/pipermail/wikitech-l/2012-February/058254.html )
[23:39:58] <ashish_d>	 aha, thanks!
[23:40:11] <RoanKattouw>	 (mailing list archives are awesome)
[23:40:30] <ashish_d>	 its now time for drafting the application though :-)
[23:40:35] <RoanKattouw>	 Yes
[23:40:40] <RoanKattouw>	 Are you gonna be doing this on a wiki page?
[23:40:47] <ashish_d>	 Yes
[23:40:53] <RoanKattouw>	 Cool
[23:43:32] <RoanKattouw>	 Send an e-mail to Trevor, Gabriel and me when you've got something up, and we'll look at it
[23:43:51] <RoanKattouw>	 (Well at least I will -- I can't make any promises for Trevor, he's gonna be drugged up on painkillers for a while)
[23:44:02] <sumanah>	 RoanKattouw: (or wikitext-l, right?)
[23:44:08] <RoanKattouw>	 Yes, that too
[23:44:09] <ashish_d>	 Yes, very soon!
[23:44:19] <RoanKattouw>	 Well -- wikitext-l is kind of the parser list really
[23:44:28] <sumanah>	 no reason to do most GSoC discussion in private when .... it's VE too
[23:44:28] <ashish_d>	 wikitech-l?
[23:44:31] <sumanah>	 no
[23:44:37] <sumanah>	 well, that would be a fine choice as well
[23:44:43] <RoanKattouw>	 Yeah
[23:44:46] <sumanah>	 https://lists.wikimedia.org/mailman/listinfo/wikitext-l
[23:48:10] <sumanah>	 ashish_d: btw, glad to see you again, sorry I haven't been around on IRC more.
[23:48:16] <sumanah>	 Lots of in-person meetings the last 2 weeks.
[23:49:27] <ashish_d>	 sumanah: thats okay, even I've been pretty much dormant around, been juggling with things at school :p
[23:50:02] <sumanah>	 ashish_d: I am trying to make a "dorm" "dormant" pun and failing.
[23:50:05] <sumanah>	 Just so you know. :D
[23:50:33] <ashish_d>	 hehe
[23:51:07] <ashish_d>	 dormed and dormant at school, how does that sound?
[23:52:40] <sumanah>	 sure :)
[23:53:36] <RoanKattouw>	 "dorm" and "dormant" very likely have the same root (French "dormir" or whatever the Latin precursor of that was)
[23:53:40] <RoanKattouw>	 So in my book that doesn't count :)
[23:53:46] * JRWR  throws the Mediawiki install disc across the room!

[23:53:47] <sumanah>	 like "habeas corpse"
[23:53:48] <sumanah>	 yeah
[23:53:51] <JRWR>	 I hate session cookies
[23:53:55] <sumanah>	 JRWR: my sympathies
[23:54:21] <JRWR>	 cant cache a damn thing if my install of mw wants to set a session cookie for every pag
[23:54:28] <JRWR>	 anon user or not
[23:54:52] <RoanKattouw>	 I'm not sure if that can be disabled or not
[23:55:16] <RoanKattouw>	 Anyway the way WMF does it is by varying on a specific pattern in the Cookie header
[23:58:08] <JRWR>	 but its calling set-cookie on any page
[23:58:19] <JRWR>	 that throws off nginx fastcgi cache
[23:58:56] <JRWR>	 example headers
[23:58:57] <JRWR>	 http://paste.ubuntu.com/884114/