[03:38:13] <TimStarling>	 AaronSchulz: still here?
[03:38:44] <TimStarling>	 <binasher> maybe unrelated but we're logging a lot of
[03:38:45] <TimStarling>	 <binasher> Exception from line 94 of /usr/local/apache/common-local/php-1.21wmf6/extensions/ConfirmEdit/FancyCaptcha.class.php: Ran out of captcha images
[10:36:52] <Nemo_bis>	 hmpf, so many interns blogs and none of them can be added to the planet
[12:08:01] <gerrit-wm>	 New review: Hashar; "Yeah I am evil. integration/* git repositories are configured to require fast forward commits :-D" [integration/zuul-config] (master); V: 0 C: 0;  - https://gerrit.wikimedia.org/r/42583
[12:11:26] <gerrit-wm>	 New patchset: Hashar; "mwext-EventLogging-pep8 is now blocking on failure" [integration/zuul-config] (master) - https://gerrit.wikimedia.org/r/42583
[12:12:06] <gerrit-wm>	 New review: Hashar; "Merging per Ori approval" [integration/zuul-config] (master); V: 2 C: 2;  - https://gerrit.wikimedia.org/r/42583
[12:12:07] <gerrit-wm>	 Change merged: Hashar; [integration/zuul-config] (master) - https://gerrit.wikimedia.org/r/42583
[12:12:54] <gerrit-wm>	 New review: Hashar; "Deployed" [integration/zuul-config] (master) - https://gerrit.wikimedia.org/r/42583
[13:09:21] <hashar>	 hello
[13:40:47] <gerrit-wm>	 New patchset: Stefan.petrea; "Adding new pageview reports mobile (in progress)" [analytics/wikistats] (master) - https://gerrit.wikimedia.org/r/41979
[14:06:08] <hashar>	 ^demon: would you be available to abandon a few changes for me please ? :-D
[14:06:34] <^demon>	 Sure, links?
[14:07:03] <hashar>	 ^demon: https://gerrit.wikimedia.org/r/#/c/27937/1  some debug code for twn.net
[14:07:30] <hashar>	 ^demon: that is not needed in core nor still used on twn afaik
[14:07:49] <^demon>	 Yeah, and if twn needs it they'd be cherry picking it anyway.
[14:08:00] <hashar>	 ^demon: https://gerrit.wikimedia.org/r/#/c/21897/ some refactoring code that introduce bugs and is unneeded imo.
[14:09:23] <hashar>	 annnnd https://gerrit.wikimedia.org/r/#/c/9598/  which attempt to add memcached support to Language::fetchLanguageNames() which is like … unneeded since we already have faster alternative (aka pure PHP code without any db access).
[14:09:42] <hashar>	 (I am attempting to clean up my Gerrit dashboard)
[14:14:20] <hashar>	 ^demon: thanks sir :-]
[14:14:25] <^demon>	 yw.
[14:15:46] <hashar>	 Reedy: still alive ? :-D
[14:15:59] <hashar>	 Reedy: I got the wmf6 and wmf7 branches cloned on the beta shared directory
[14:17:24] <hashar>	 we have to pick up beta wikis and switch them from master to one of the wmf branch, not sure which one to pick up though
[14:17:25] <hashar>	  http://deployment.wikimedia.beta.wmflabs.org/wiki/Special:SiteMatrix
[14:18:30] <^demon>	 I'd use 1.21wmf7.
[14:18:49] <^demon>	 beta should probably use the latest of the two deployed versions.
[14:19:32] <hashar>	 ^demon: that is to test out git deploy, so we need wikis in both wmf branches
[14:20:31] <hashar>	 I will just pick to english projects
[14:21:02] <^demon>	 Ah, gotcha.
[14:29:09] <hashar>	 Reedy: https://gerrit.wikimedia.org/r/#/c/42951/
[14:31:12] <hashar>	 (Can't contact the database server: Access denied for user ''@'localhost' to database 'my_wiki' (localhost))
[14:31:14] <hashar>	 mouahahha
[14:36:27] <hashar>	 I am doomed
[14:36:38] <hashar>	 beta now complains about LocalSettings.php not found hehe
[14:37:36] <sumanah>	 what!
[14:37:41] <sumanah>	 hi physikerwelt, how are you?
[15:38:56] <Platonides>	 hashar?
[15:39:07] <hashar>	 Platonides: hello :-]
[15:39:52] <Platonides>	 why can't you simply approve all of https://gerrit.wikimedia.org/r/#/q/status:open+project:mediawiki/tools/code-utils+owner:platonides,n,z ? :P
[15:40:48] <Platonides>	 about dependencies, I prefer a linear history when suitable
[15:41:52] <Platonides>	 although a long list of changes has piled there
[15:42:07] <^demon>	 For linear history--we could enable ff-only for code-utils.
[15:42:15] <hashar>	 yeah lets do that
[15:42:47] <^demon>	 done.
[15:42:49] <hashar>	 Platonides: with ff-only configured in Gerrit for the mw/tools/code-utils.git repository, people are forced to rebase to latest master to get Gerrit to merge the change.
[15:42:52] <hashar>	 ^demon: \O/
[15:47:46] <hashar>	 Platonides: and for https://gerrit.wikimedia.org/r/#/q/status:open+project:mediawiki/tools/code-utils+owner:platonides,n,z
[15:48:00] <hashar>	 Platonides: I would actually have to review the change properly :-]
[15:49:13] <Platonides>	 I agree it's not a trivial script...
[15:49:25] <Platonides>	 but who else could I bug about it? :)
[15:50:18] <Platonides>	 you rejected to review a trivial fix removing a backslash today, however
[15:50:18] <Platonides>	 so that must mean you don't like too-ease patches either xD
[15:50:34] <Platonides>	 * too-easy
[15:51:04] <hashar>	 possibly
[16:05:26] <hashar>	 I am out, got two beta wikis to run out of wmf branches :-]
[16:05:29] * hashar  waves

[16:10:00] <sumanah>	 Platonides: an OPW intern who is moving to Spain is curious about the open source scene in Spain.  Like, what conferences to go to, whom to contact, what specific communities are particularly active
[16:10:15] <sumanah>	 any suggestions? #opw on the GNOME IRC network (GIMPnet) is where we're talking now
[16:24:28] <Dereckson>	 Platonides: do you see an obstacle to use $arguments = escapeshellcmd( implode( ' ', $argv )  ) ; system( 'env php -d enable_dl=On ' . $arguments, $returnCode ); or should each argument be escaped with escapeshellarg?
[16:24:45] <Dereckson>	 (re: https://gerrit.wikimedia.org/r/#/c/42952/1/check-vars.php)
[16:27:38] <Dereckson>	 ^demon: Gerrit doesn't seem to advertise correctly chmod changes
[16:28:25] <^demon>	 example? I seem to remember it did.
[16:28:35] <Dereckson>	 ^demon: see https://gerrit.wikimedia.org/r/#/c/42950/ - stylize.php has no content change, so chmod is advertised
[16:28:43] <Dereckson>	 for every other files chmoded, no mention
[16:29:05] <^demon>	 Ah, gotcha.
[16:30:04] <^demon>	 Possibly https://code.google.com/p/gerrit/issues/detail?id=1236?
[16:30:13] <^demon>	 1236 seems vague though.
[16:30:55] <Dereckson>	 Indeed.
[16:52:02] <sumanah>	 multichill: ping re Amsterdam hackathon - can you please close the straw poll and announce a date?
[17:24:08] <sumanah>	 DanielK_WMDE__: btw not sure whether you saw my mention of http://www.crummy.com/2012/12/21/1 in case you want to betaread Leonard's next book on RESTful web API design -- it would be several months from now
[17:35:16] <marktraceur>	 Ooh, ooh, pick me too
[17:37:46] <marktraceur>	 sumanah: I'm about halfway through the first one that ori-l kindly loaned me, in response to the bug I had opened about making MW more RESTful
[17:38:24] <marktraceur>	 (so it may interest ori-l to see the above also)
[18:19:54] <DanielK_WMDE__>	 sumanah: thanks for the link, will have a look!
[18:20:46] <multichill>	 sumanah: Will do, probably later this week
[18:21:23] <sumanah>	 Thanks multichill -- am I right to say that it's a 90% chance that May 25th is the weekend we'll have it?
[18:21:34] <multichill>	 Make that 99%
[18:28:43] <sumanah>	 chrismcmahon: "So the example for Search above (which is in fact already automated)" -- when you say it "is in fact already automated" I'm going to point to https://github.com/wikimedia/qa-browsertests/blob/master/features/search.feature
[18:29:06] <chrismcmahon>	 sumanah: lovely, thanks
[18:31:06] <sumanah>	 chrismcmahon: (it was confusing because I could tell whether "is automated" was an adjective or a verb, really)
[18:31:26] <sumanah>	 Improved to So the example description we gave above, for Search, becomes [https://github.com/wikimedia/qa-browsertests/blob/master/features/search.feature this automated test]:
[18:31:33] <chrismcmahon>	 sumanah: I do appreciate the editing very much!
[18:31:43] <sumanah>	 cool
[18:31:52] <sumanah>	 er, I mean, I *couldn't* tell. Speaking of editing!
[18:31:58] <chrismcmahon>	 :)
[18:32:15] <sumanah>	 qgil: chrismcmahon - I look forward to seeing you playtest this documentation with a member of your target audience.
[18:32:39] <sumanah>	 That helped me immensely whenever I wrote or ran a tutorial!
[18:33:21] <sumanah>	 https://www.mediawiki.org/wiki/QA/How_to_contribute_to_browser_testing is now a lot more usable.
[18:35:30] <gerrit-wm>	 New patchset: Hashar; "whitelist Marius Hoch (Hoo Man)" [integration/zuul-config] (master) - https://gerrit.wikimedia.org/r/39711
[18:35:56] <gerrit-wm>	 New review: Hashar; "rebased, fixed conflict.  Deploying :-)" [integration/zuul-config] (master); V: 2 C: 2;  - https://gerrit.wikimedia.org/r/39711
[18:35:57] <gerrit-wm>	 Change merged: Hashar; [integration/zuul-config] (master) - https://gerrit.wikimedia.org/r/39711
[18:36:05] <qgil>	 sumanah, I saw your email about a potential candidate
[18:36:39] <preilly>	 ori-l: ping
[18:36:58] <Reedy>	 ori-l: Now you've done it.
[18:37:33] <preilly>	 Reedy: ha ha ha
[18:40:17] <qgil>	 sumanah, you said you configure the IRC client to suggest the nick of the last perso you talked to instead of alphabetic? Can't find it in XChat...
[18:40:58] <sumanah>	 qgil: - Settings, Preferences, Interface, Userlist
[18:41:09] <sumanah>	 do you see it?
[18:41:58] <qgil>	 sumanah, I'm there but can't find the option
[18:43:37] <sumanah>	 qgil: what version XChat are you using?
[18:43:55] <qgil>	 sumanah, XChat 2.8.8
[18:44:10] <sumanah>	 I see a "User list sorted by:" dropdown list of 5 options, also XChat 2.8.8
[18:45:13] <qgil>	 sumanah, but I thought you meant that when I start typing "sum" + tab I would get sumanah suggested instead of summum or sumatra if I just chatted with sumanah
[18:45:46] <sumanah>	 oh. I'm sorry .... it's *last spoke*, not *last spoke with*
[18:45:50] <sumanah>	 :/
[18:46:16] <qgil>	 sumanah, still, I don't find any option for last spoke
[18:46:21] <sumanah>	 I do
[18:46:37] <sumanah>	 we may need to Hangout and you screenshare!
[18:46:58] <qgil>	 sumanah, boh - next time.  :)  Sorry for taking your time.
[18:47:03] <sumanah>	 never mind!
[18:47:05] <sumanah>	 sorry!
[18:47:10] <sumanah>	 I meant Input Box
[18:47:20] <sumanah>	 Preferences, Interface, Input Box -- "nick completion" section
[18:47:30] <qgil>	 sumanah, ah, there. Fixed now.  :)
[18:47:32] <sumanah>	 I'm sorry, I was distracted and was looking at the wrong thing (I was in a meeting)
[18:47:35] <sumanah>	 (am still)
[18:47:43] <qgil>	 thankk yoou
[18:47:55] * sumanah  is frustrated with herself for speaking wrongly at first!

[18:56:32] * preilly  gives sumanah a hug

[18:56:50] * sumanah  appreciates the support

[18:56:51] * preilly  and thinks don't be frustrated...

[19:12:04] <preilly>	 ^demon|lunch: ping
[19:12:09] <^demon|lunch>	 Hi
[19:12:48] <preilly>	 ^demon: I've got a git related question
[19:13:03] <^demon>	 Shoot.
[19:13:03] <preilly>	 ^demon: do you have a second?
[19:13:31] <preilly>	 ^demon: I need to place https://github.com/wikimedia/WikipediaMobileFirefoxOS in gerrit.wikimedia.org:29418/operations/mediawiki-config/docroot/bits
[19:13:49] <preilly>	 ^demon: what is the best way to do that in your opinion?
[19:14:18] <^demon>	 submodule?
[19:14:33] <preilly>	 ^demon: to a github repo?
[19:14:51] <^demon>	 I don't see why not?
[19:14:55] <^demon>	 It's one we maintain.
[19:15:07] <^demon>	 Although I suppose we could move the repo into gerrit.
[19:15:40] <preilly>	 ^demon: okay so you are cool with it being a submodule?
[19:16:03] <^demon>	 Yeah, submodules are fine. Might want to check with $otherPeople though.
[19:16:09] <^demon>	 wmf-config is changing drastically for git-deploy.
[19:16:14] <^demon>	 So just wanna make sure.
[19:16:23] <preilly>	 ^demon: that was why I was asking you. ;-)
[19:16:42] <^demon>	 I've not done anything for git-deploy, so I dunno that :)
[19:17:08] <preilly>	 ^demon: okay thanks
[19:18:20] <MaxSem>	 Nikerabbit, PHP Warning:  Missing argument 3 for SolrTTMServer::createDocument(), called in /usr/local/apache/common-local/php-1.21wmf7/extensions/Translate/ttmserver/SolrTTMServer.php on line 159 and defined in /usr/local/apache/common-local/php-1.21wmf7/extensions/Translate/ttmserver/SolrTTMServer.php on line 179
[19:20:02] <Krinkle>	 hasharAway: ping
[19:20:23] <hashar>	 Krinkle: pong somehow
[19:20:47] <Krinkle>	 hashar: 2 things from yesterday: 1) plan/progress on vagrant/vm, 2) how/where to get zuul to set jenkins to read checkstyle files?
[19:22:32] <hashar>	 1) no plan. I am busy with beta till the end of february.  Though I will try to at least have a look at it, write something about it or get ori to help us on vagrant :)
[19:22:48] <Krinkle>	 ok
[19:23:24] <hashar>	 Krinkle: for 2) check the publisher pep8 macro in jenkins-job-builder-config.git  macros.yaml
[19:23:38] <Krinkle>	 ok
[19:23:57] <hashar>	 Krinkle: the upstream doc is at http://ci.openstack.org/jenkins-job-builder/publishers.html
[19:24:03] <Krinkle>	 hashar: I'm going to move jshint back from grunt to bash. Unnececary abstraction and complication.
[19:24:24] <hashar>	 Krinkle: basically copy paste, replace pep8 with one of the valid systems: checkstyle, codenarc, cpd, cpplint, csslint, findbugs, fxcop, gendarme, jcreport, jslint, pep8, pmd, pylint, simian, stylecop
[19:24:29] <Krinkle>	 okay
[19:24:47] <Krinkle>	 hashar: so the plan for qunit is as follows
[19:24:49] <hashar>	 then we will be able to reuse that publishing macro on a job
[19:26:00] <Krinkle>	 hashar: since the runs are synchronous, we don't need to snapshot. We can just have a temporary symlink in a /dir/to/some/www/uniqueid, run the installer and run grunt qunit (uses phantomjs)
[19:26:15] <Krinkle>	 I'll also have to add a virtualhost for it and a deny rule
[19:26:16] <Krinkle>	 sounds right?
[19:26:44] <preilly>	 Reedy: Can you please review https://gerrit.wikimedia.org/r/#/c/43009/
[19:27:50] <hashar>	 Krinkle: or just make the virtual host to point to /var/lib/jenkins/jobs/mediawiki-core-qunit/workspace :-D
[19:28:17] <Krinkle>	 That wouldn't scale for other repos, and hardcodes the job name
[19:28:41] <Krinkle>	 also, we need a unique url for each run
[19:28:52] <Krinkle>	 to avoid cache issues (just in case)
[19:28:59] <preilly>	 AaronSchulz: take a look at https://gerrit.wikimedia.org/r/#/c/43009/1 please
[19:29:30] <Krinkle>	 hashar: btw, how would I go about installing a gem? puppet?
[19:29:37] <Krinkle>	 (we need jsduck)
[19:30:22] <hashar>	 isnt jsduck to generate documentation?
[19:30:26] <Krinkle>	 Yes
[19:30:36] <hashar>	 I guess it is lower priority than phantomJS :-]
[19:30:47] <Krinkle>	 I want to (after jshint/qunit) auto-generate it from post-merge on visualeditor
[19:30:48] <hashar>	 anyway, debian most probably has a tool to convert a ruby gem to a debian package
[19:31:04] <Krinkle>	 Can we not do that?
[19:31:07] <hashar>	 we could then upload the resulting .deb on apt.wm.org and make it available on the server :-D
[19:31:12] <hashar>	 ask ops :-]
[19:31:19] <Reedy>	 preilly: Looks fine bar jenkins complaining
[19:31:20] <hashar>	 but I am pretty sure they will want a package
[19:31:21] <AaronSchulz>	 preilly: why don't you use https://github.com/WikipediaMobileFirefoxOS.git btw?
[19:31:42] <AaronSchulz>	 instead of a url that gives 301
[19:31:47] <Krinkle>	 hashar: I can checkout the code in var/lib/git and just reference the bin ;-). It isn't compiled or installed...
[19:31:49] <Reedy>	 heh
[19:31:51] <preilly>	 AaronSchulz: I fixed the issue and jenkins is good now https://gerrit.wikimedia.org/r/#/c/43009/1
[19:31:54] <hashar>	 Krinkle: OR we do the same madness we did for the npm modules and put the gems in integration/jenkins.git under some dir like  /gems . Then have ruby to look up in that path
[19:32:15] <preilly>	 AaronSchulz: I changed it to git://github.com/wikimedia/WikipediaMobileFirefoxOS.git
[19:32:20] <preilly>	 for Git Read-Only
[19:32:36] <AaronSchulz>	 ok, nice
[19:32:45] <Krinkle>	 hashar: Hm..
[19:33:11] <preilly>	 AaronSchulz, Reedy: can one of you +2 it or whatever the proper next step is ;-)
[19:33:17] <Reedy>	 I just did
[19:33:26] <Reedy>	 Didn't submit it as wasn't sure when you're wanting to deploy it
[19:33:48] <preilly>	 Reedy: I'd like to deploy it as soon as possible
[19:36:03] <ori-l>	 ahoy
[19:36:21] <ori-l>	 Reedy / preilly: I saw mentions of my nick but my scroll back lost the context
[19:37:08] <preilly>	 ori-l: give me a sec
[19:37:13] <Nikerabbit>	 MaxSem: that's already fixed in master
[19:37:18] <ori-l>	 preilly: np.
[19:37:43] <MaxSem>	 Nikerabbit, it's currently on top of fatalmonitor
[19:43:13] <ori-l>	 [09:38:24] <marktraceur:#wikimedia-dev> (so it may interest ori-l to see the above also)
[19:43:24] <ori-l>	 marktraceur: i lost the context for that too, sadly.
[19:43:28] <Krinkle>	 hashar: Is there a tool you can recommend? I see "fpm" and "gem2deb" are popular.
[19:43:34] * ori-l  goes to tinker with his znc configs.

[19:43:53] <marktraceur>	 ori-l: Leonard is writing another REST book and wants beta testers!
[19:43:55] <Krinkle>	 http://blog.copyninja.info/2011/05/tutorial-packaging-ruby-gem-in-debian.html http://stackoverflow.com/questions/7116377/create-a-debian-package-from-a-ruby-gem https://github.com/jordansissel/fpm/wiki/ConvertingGems
[19:44:39] <hashar>	 Krinkle: I have ZERO idea about it sorry :/
[19:44:42] <Krinkle>	 ok
[19:44:54] <ori-l>	 Krinkle: what are you looking to package?
[19:44:59] <hashar>	 would have to as the debian maintainer for gem pacages
[19:45:01] <Krinkle>	 ori-l: a gem
[19:45:04] <hashar>	 or ori-l :-]]]]]]]
[19:45:06] <ori-l>	 Krinkle: right, but what?
[19:45:13] <Krinkle>	 ori-l: `gem install jsduck`
[19:45:25] <Krinkle>	 https://github.com/senchalabs/jsduck
[19:45:45] <Krinkle>	 Basically make it so that I can give ops a thing they put on apt.wikimedia.org so we can install it via puppet on a server.
[19:46:03] <hashar>	 Krinkle: is that needed for phantomjs / quint ?
[19:46:12] <hashar>	 Krinkle: or do you want to get jsduck first ?
[19:46:29] <Krinkle>	 Not needed, I just want to research it first because that is a variable.
[19:46:35] <Krinkle>	 And there are people here now.
[19:46:58] <Krinkle>	 the other two things (qunit/phantomjs) I've done times and times before. I'll do that monkey work later today.
[19:47:06] <Krinkle>	 But you (plural) are here now :)
[19:47:24] <hashar>	 got it :-]
[19:47:39] <Krinkle>	 more https://we.riseup.net/debian/debian-packaging-of-rubygems
[19:47:50] <Krinkle>	 prff. I have no idea. Never created a deb before.
[19:48:21] <hashar>	 Krinkle: also note we have ruby 1.8 on gallium
[19:48:32] <Krinkle>	 hashar: We'll have to pumpt 1.9 on it
[19:48:35] <hashar>	 so if JSDuck is using ruby 1.9 we are in trouble
[19:48:53] <Krinkle>	 It is in debian, so shouldn't be a problem.
[19:49:04] <Krinkle>	 We don't use ruby for anything there, no issues I assume.
[19:49:09] <ori-l>	 Krinkle: other alternative is to fork it to gerrit -- there are puppet modules that load modules that way
[19:49:17] <hashar>	 Krinkle: we use for puppet at least :-)
[19:49:29] <hashar>	 Krinkle: and I don't think puppet 2.7 runs with 1.9.
[19:49:36] <hashar>	 an issue you might want to raise to ops list
[19:49:39] <Krinkle>	 ori-l: how does forking help?
[19:49:44] <Krinkle>	 It is still going to be a gem.
[19:50:11] <Krinkle>	 it should work as Package[] / apt-get install
[19:50:51] <AaronSchulz>	 preilly: every time I see weird google results and then notice the "showing results for X" thing I want to flip a shit
[19:51:10] <Krinkle>	 hashar: actually, it works fine on 1.8, that's what we have on labs
[19:51:12] <preilly>	 AaronSchulz: ha ha
[19:51:21] <Krinkle>	 integration-apache1:~$ ruby --version
[19:51:21] <Krinkle>	 ruby 1.8.7 (2011-06-30 patchlevel 352) [x86_64-linux]
[19:52:05] <Krinkle>	 Yep, same on gallium
[19:53:20] <hashar>	 Krinkle: that is the ruby version that comes with Ubuntu Precise.
[19:53:28] <hashar>	 \O/
[19:54:59] <AaronSchulz>	 preilly: https://github.com/nicolasff/phpredis/issues/166 :(
[19:58:13] <preilly>	 AaronSchulz: oh damn that sucks
[19:58:24] <preilly>	 AaronSchulz: you should chime in on that issue on github
[19:58:38] <preilly>	 AaronSchulz: it could also be a fun C project for you ;-)
[20:03:47] <AaronSchulz>	 preilly: what is the version of the redis server we use?
[20:03:51] <AaronSchulz>	 is it 2.6?
[20:04:15] <preilly>	 AaronSchulz: I think so
[20:04:24] <preilly>	 AaronSchulz: I'd need to double check
[20:05:50] <preilly>	 AaronSchulz: 2:2.6.3-wmf1
[20:06:05] <AaronSchulz>	 ok
[21:41:48] <Nemo_bis>	 jdlrobson: silly question on https://gerrit.wikimedia.org/r/#/c/43001/ , on mobile where? a beta wiki already running this patch?
[21:42:04] <jdlrobson>	 dev version Nemo_bis
[21:42:15] <jdlrobson>	 if you run a local instance and checkout that patch
[21:42:38] <jdlrobson>	 (you also need to enable the alpha mode via the MobileOptions special page)
[21:42:42] <Nemo_bis>	 ah
[21:43:05] <Nemo_bis>	 well, I'll test it together with all the folks then
[21:44:36] <jdlrobson>	 thanks Nemo_bis :)
[21:44:48] <jdlrobson>	 Nemo_bis: not urgent though :)
[22:07:44] <preilly>	 ori-l: ping (2)
[22:31:41] <preilly>	 robla: do you have a link to the gerrit change that includes an ORM?
[22:32:00] <preilly>	 StevenW: thanks, for the link to https://meta.wikimedia.org/wiki/Research:Account_creation_UX/CAPTCHA
[22:32:05] <preilly>	 StevenW: that helped
[22:32:34] <robla>	 preilly: just about to head to a meeting, but the ORM stuff is already checked in (has been for a while)
[22:34:01] <preilly>	 robla: Okay thanks
[22:34:08] <preilly>	 robla: but in an extension right?
[22:34:16] * preilly  really hopes that the case

[22:34:29] <preilly>	 s/the/is the/
[22:37:20] <StevenW>	 preilly: of course
[22:37:46] <StevenW>	 experiments without some kind of documentation of their purpose should always be challenged I think
[22:40:40] <preilly>	 StevenW: indeed
[22:42:36] <binasher>	 preilly: are you talking about luke's new orm?
[22:42:42] <binasher>	 er not luke
[22:42:44] <binasher>	 mathias
[22:42:48] <preilly>	 binasher: yes
[22:43:19] <binasher>	 i think robla is confusing it for something else
[22:43:23] <TimStarling>	 bots will discover the change instantly
[22:43:52] <TimStarling>	 but I suppose it would be too hard to classify account creations into bots and humans
[22:43:55] <binasher>	 there was an orm that was merged in months back with discussion on wikitech. robla: a second one is being written, seemingly in a vacuum
[22:44:54] <TimStarling>	 binasher: jeroen's "ORM" was not a large thing
[22:45:17] <binasher>	 TimStarling: agreed. i think they'll also have trouble quantifying quality edits, assuming there's an edit increase in that time. if lots of new humans sign up and make meaningful edits, plenty will get reverted as well
[22:45:24] <TimStarling>	 so it doesn't surprise me that someone might find that it doesn't fit all their needs
[22:45:42] <StevenW>	 Tim: you think so? Everyone else has been insisting they won't.
[22:45:55] <TimStarling>	 I don't know why they would say that
[22:46:19] <binasher>	 TimStarling: have you reviewed mathias's?
[22:46:38] <TimStarling>	 StevenW: spambots work by scraping the HTML of whatever page they happen to end up on, identifying form fields by labels, etc.
[22:47:03] <TimStarling>	 if there's an email address field then they'll put a valid email address in it
[22:47:08] <TimStarling>	 if there's a captcha field, they will solve it
[22:47:20] <StevenW>	 The other argument I hear is that even if would identify it immediately, it doesn't matter because we won't feel the full effect until spammers learn that we don't have a captcha to beat on registration.
[22:47:27] <StevenW>	 Which is an argument I buy more.
[22:47:38] <TimStarling>	 they generally don't know that they are on a mediawiki instance, they are just generic spambots for all sorts of user contributed content
[22:48:16] <StevenW>	 In any case, if spam doesn't explode in that two hour window, we're not going to shout from the rooftops that it's evidence that the thing is totally useless.
[22:48:42] <TimStarling>	 we see a lot of spam which is not correctly formatted for wikitext
[22:48:55] <StevenW>	 as in raw HTML?
[22:49:01] <TimStarling>	 they usually just assume filtered HTML, like most blog comments
[22:49:26] <StevenW>	 Gotcha
[22:49:40] <TimStarling>	 so I think it's unlikely that the presence of a captcha on wikipedia is stopping the spambots from visiting it
[22:50:09] <TimStarling>	 the operators probably aren't even aware of what sites they are spamming, they just generate target lists automatically, by crawling
[22:50:43] <TimStarling>	 binasher: no
[22:52:11] <binasher>	 hell, we get lots of spam submitted as search queries
[23:05:37] <preilly>	 AaronSchulz: https://www.mediawiki.org/wiki/Extension:EventLogging/Schemas
[23:05:46] <AaronSchulz>	 preilly: thanks
[23:08:49] <preilly>	 AaronSchulz: thanks for taking a look
[23:23:43] <spagewmf>	 preilly you are talking about three unrelated things.  Jeroen De Dauw's ORMTable in core, Matthias Mullie's new DataModel extension (for AFT and eventually other Features), and E3's use of JSON schemas to define the events we log.
[23:23:43] <robla>	 marktraceur: etherpad on wmflabs is down
[23:23:52] <robla>	 nm
[23:25:29] <preilly>	 spagewmf: please don't try to suppose what I'm talking about
[23:26:02] <preilly>	 spagewmf: also I'm well aware of what I'm talking about and looking at ;-)
[23:26:38] <marktraceur>	 robla: I disagree
[23:26:42] <marktraceur>	 Oh, K
[23:28:17] <spagewmf>	 preilly, glad to hear it :)
[23:33:21] <preilly>	 spagewmf: The idea of having to do things like:  $sample = new DataModelSample; seem really bad to me
[23:34:05] <preilly>	 spagewmf: but things like http://schema.org/ make total sense to me
[23:35:09] <preilly>	 spagewmf: and Facebook Open Graph as well
[23:37:36] <binasher>	 why is the functionality that DataModel is striving for being implemented by the features team, in an extension, and not in cooperation with platform or explicitly by AaronSchulz and TimStarling in core?
[23:56:41] <preilly>	 AaronSchulz: Have you ever looked at http://codex.wordpress.org/Using_Custom_Fields
[23:57:58] <AaronSchulz>	 I don't think so, I've seen the wp schema though