[00:55:12] <marktraceur>	 gwicke_: Hey, busy tonight?
[00:56:24] <gwicke_>	 marktraceur: am busy with not being busy for a change ;)
[00:56:50] <marktraceur>	 gwicke_: Care to join me in my regular Sunday insanity? http://rose.makesad.us/pipermail/beowulf-cluster/2012-December/000009.html
[00:57:34] <gwicke_>	 beowulf? are you involved in that project?
[00:57:49] <marktraceur>	 gwicke_: It's not a project, just a meeting of free software people :)
[00:58:04] <marktraceur>	 It's a spin-off of the meeting in Boston at a restaurant called Grendel's
[00:58:14] <marktraceur>	 (we're *that* nerdy)
[00:58:43] <gwicke_>	 interesting- beowulf used to be a low-cost HPC clustering project
[00:58:59] <gwicke_>	 linux-based normally
[00:59:02] <marktraceur>	 Yes, that's one of the wordplays involved
[00:59:30] <marktraceur>	 All of us being free software people, most of us remembering that project (myself not among them, but I've had it explained to me)
[01:00:10] <gwicke_>	 k, its heydays are probably over by now
[01:00:30] <gwicke_>	 I'm more in clean-up and reading mood tonight
[01:00:35] <marktraceur>	 AOK!
[01:00:52] <gwicke_>	 but sounds like fun!
[01:01:10] <ori-l>	 gwicke_: beowulf's heyday was over in the 11th century or thereabouts :P
[01:01:20] <gwicke_>	 hehe- yes
[01:01:36] <gwicke_>	 read up on that when I came across the HPC beowulf for the first time
[01:01:49] <marktraceur>	 ori-l: When he killed Grendel (wordplay #2)
[01:02:28] <gwicke_>	 this weekend has been the usual recovery shutdown experience..
[01:03:15] <ori-l>	 how are things with parsoid?
[01:03:37] <gwicke_>	 pretty good- we fared better than expected
[01:04:02] <gwicke_>	 clean diffs for the most part make happy early adopters
[01:05:23] <ori-l>	 i've noticed a fever of puppet activity around parsoid -- is it exposed to the mediawiki boxen as a network RPC?
[01:05:48] <gwicke_>	 yes, a stateless HTTP service
[01:06:16] <gwicke_>	 the load balancing now actually works, so we don't run on a single machine any more ;)
[01:06:31] <ori-l>	 how did you implement load balancing?
[01:06:40] <ori-l>	 node's cluster module?
[01:06:42] <gwicke_>	 Roan did all that- he used LVS
[01:06:59] <gwicke_>	 and each box uses cluster to manage workers
[01:07:17] <gwicke_>	 we start number of cores + 3 workers
[01:07:32] <gwicke_>	 but each worker can only handle a single request
[01:07:37] <gwicke_>	 at a time
[01:08:24] <ori-l>	 are you planning to keep this basic architecture, or to bundle everything into core?
[01:09:10] <gwicke_>	 ori-l: the latest idea is to eliminate wikitext parsing from MW in the long run ;)
[01:09:11] <ori-l>	 it'd be nice if our architecture was more message-oriented rather than monolithic. that would allow for greater diversity of languages and tools. whatever's right for the job.
[01:09:36] <gwicke_>	 following the idea of no parser being the fastest parser
[01:10:12] <ori-l>	 NoParsing, à la NoSQL
[01:10:18] <gwicke_>	 lot of things to research before making a plan on that though
[01:10:45] <gwicke_>	 and discussion obviously
[01:10:50] <ori-l>	 yes, otherwise we'd end up with a NoWiki :)
[01:11:27] <gwicke_>	 we can still have Parsoid as a text UI
[01:11:36] <ori-l>	 have you used erlang and OTP, btw?
[01:11:59] <gwicke_>	 ori-l: played a little bit with it, but not much
[01:12:03] <gwicke_>	 mostly in college
[01:12:44] <gwicke_>	 spent more time with Haskell
[01:13:13] <ori-l>	 haskell is ultimately too WTF for me
[01:13:42] <gwicke_>	 ori-l: there is a pretty nice book called real world haskell
[01:13:51] <gwicke_>	 http://book.realworldhaskell.org/read/
[01:14:10] <gwicke_>	 but I agree re components and message passing
[01:14:28] <gwicke_>	 I'm trying to give everything an URL and make it all pure / RESTful
[01:15:44] <ori-l>	 oh, that's neat. i'll have to toy around with it one of these days
[01:15:46] <gwicke_>	 if we can avoid spending too much time on writing a super-fast parser we might instead be able to make progress on fragment caching and incremental updates
[01:16:18] <gwicke_>	 which has more potential in the longer term
[01:16:54] <ori-l>	 i think so too
[01:18:11] <gwicke_>	 the question marks are mainly in the interaction with existing extensions
[01:18:30] <ori-l>	 hardware is cheap and with an architecture oriented around message-passing between discrete components you can scale by adding hardware
[01:18:44] <gwicke_>	 how many depend on the frame object for example
[01:19:36] <ori-l>	 i'm reading http://www.mediawiki.org/wiki/Manual:Tag_extensions now, didn't know about any of this stuff
[01:21:28] <ori-l>	 ok, back to sorting out this rat's nest of python and php :-/
[01:21:32] <ori-l>	 ttyl
[01:21:59] <gwicke_>	 to each his own rat's nest to sort out ;)
[01:22:36] <marktraceur>	 Indeed. What on earth should I have for dinner?
[01:22:42] <ori-l>	 rats!
[01:22:48] * marktraceur  is happy with the complexity of that problem, thanks very much

[01:23:37] * gwicke_  solved that problem by cutting some bread and pecorino

[01:24:34] <marktraceur>	 gwicke_: Not sure I particularly feel like staying in tonight
[01:25:11] <gwicke_>	 discovered a really nice Thai place near Japan town on Friday
[01:25:39] <marktraceur>	 Hm, bit far. But possible, with BART.
[01:25:59] <gwicke_>	 http://www.yelp.com/biz/jitlada-thai-cuisine-san-francisco
[01:26:34] <marktraceur>	 Farther than I thought....maybe not
[01:27:14] <marktraceur>	 The old standby seems OK, cheap 'n' tasty Chinese food right around the corner
[01:27:24] <marktraceur>	 Maybe followed by some cafe or something
[01:27:28] <gwicke_>	 hehe
[01:27:50] <gwicke_>	 guess which Indian restaurant is right around my corner?
[01:27:59] <marktraceur>	 I clearly need to have either more friends in the Mission or a more reliable way of reaching people for dinner plans
[01:28:09] <marktraceur>	 gwicke_: I happen to know exactly, and I'm super-jealous
[01:28:23] <gwicke_>	 ;)
[01:28:23] <marktraceur>	 Though there is the Indian pizza place just up Mission....maybe I could hit the 14....
[01:28:42] <marktraceur>	 Well, maybe. I'll go figure that out now, y'all have a good evening with your rats
[01:29:02] <gwicke_>	 ya, have a good one & see you tomorrow!
[09:55:55] <hashar>	 hello
[10:08:40] * siebrand  greets hashar 

[10:08:47] <hashar>	 :)
[10:12:36] <siebrand>	 hashar: are https://gerrit.wikimedia.org/r/#/c/38553/ and https://gerrit.wikimedia.org/r/#/c/38555/ any good?
[10:12:44] <hashar>	 siebrand: yup
[10:12:57] <hashar>	 siebrand: I am checking the first one for extension dependencies
[10:13:14] <siebrand>	 hashar: afaik non of them have any formal tests yet.
[10:13:18] <hashar>	 siebrand: I will have to do some evil copy paste in Zuul configuration next
[10:13:32] <hashar>	 might end up writing a templating system in Zuul to avoid all the mess :)
[10:14:00] <siebrand>	 hashar: Just want to have linters run, so that when people do merge crap that was pointed out to them, I can take the clue bat.
[10:14:29] <hashar>	 !g I67f92915ebca7c469116571d90b1a76a68fbf97d
[10:14:29] <wm-bot>	 https://gerrit.wikimedia.org/r/#q,I67f92915ebca7c469116571d90b1a76a68fbf97d,n,z
[10:14:35] <hashar>	 siebrand: will do that :)
[10:15:38] <hashar>	 need to setup the Wikibase extension first
[10:25:05] <gerrit-wm>	 New patchset: Hashar; "ext dependencies + Wikibase jobs" [integration/jenkins-job-builder-config] (master) - https://gerrit.wikimedia.org/r/38528
[10:25:57] <gerrit-wm>	 New review: Hashar; "PS2:" [integration/jenkins-job-builder-config] (master); V: 1 C: 2;  - https://gerrit.wikimedia.org/r/38528
[10:26:05] <gerrit-wm>	 Change merged: Hashar; [integration/jenkins-job-builder-config] (master) - https://gerrit.wikimedia.org/r/38528
[10:38:33] <aude>	 hashar: on https://gerrit.wikimedia.org/r/#/c/38528/, validator is not required but otherwise looks good
[10:38:48] <hashar>	 aude: maybe I should remove it so :)
[10:38:52] <aude>	 sure
[10:39:11] <aude>	 don't think it'd break anything by being there but
[10:39:21] <aude>	 not needed
[10:39:54] <gerrit-wm>	 New patchset: Hashar; "get rid of -install jobs for MW Extensions" [integration/zuul-config] (master) - https://gerrit.wikimedia.org/r/39030
[10:40:08] <gerrit-wm>	 Change merged: Hashar; [integration/zuul-config] (master) - https://gerrit.wikimedia.org/r/39030
[10:42:40] <gerrit-wm>	 New patchset: Hashar; "get rid of -install jobs for MW extensions" [integration/jenkins-job-builder-config] (master) - https://gerrit.wikimedia.org/r/39031
[10:43:16] <gerrit-wm>	 Change merged: Hashar; [integration/jenkins-job-builder-config] (master) - https://gerrit.wikimedia.org/r/39031
[10:50:22] <gerrit-wm>	 New patchset: Hashar; "Add Babel, CleanChanges and LocalisationUpdate to job builder" [integration/jenkins-job-builder-config] (master) - https://gerrit.wikimedia.org/r/38553
[10:50:28] <gerrit-wm>	 Change merged: Hashar; [integration/jenkins-job-builder-config] (master) - https://gerrit.wikimedia.org/r/38553
[10:56:45] <gerrit-wm>	 New patchset: Hashar; "Triggers for Babel, CleanChanges and LocalisationUpdate" [integration/zuul-config] (master) - https://gerrit.wikimedia.org/r/39033
[10:56:58] <gerrit-wm>	 Change merged: Hashar; [integration/zuul-config] (master) - https://gerrit.wikimedia.org/r/39033
[10:57:57] <gerrit-wm>	 New review: Hashar; "Zuul configuration deployed with https://gerrit.wikimedia.org/r/39033" [integration/jenkins-job-builder-config] (master) - https://gerrit.wikimedia.org/r/38553
[13:04:05] <gerrit-wm>	 New patchset: Hashar; "job template for Python pep8 runs" [integration/jenkins-job-builder-config] (master) - https://gerrit.wikimedia.org/r/39044
[13:04:05] <gerrit-wm>	 New patchset: Hashar; "pep8 job for ops/puppet" [integration/jenkins-job-builder-config] (master) - https://gerrit.wikimedia.org/r/39045
[13:09:27] <gerrit-wm>	 New patchset: Hashar; "pep8 jobs are non voting (for now)" [integration/zuul-config] (master) - https://gerrit.wikimedia.org/r/39046
[13:09:27] <gerrit-wm>	 New patchset: Hashar; "lint ops/puppet python files (non voting)" [integration/zuul-config] (master) - https://gerrit.wikimedia.org/r/39047
[13:09:49] <gerrit-wm>	 Change merged: Hashar; [integration/jenkins-job-builder-config] (master) - https://gerrit.wikimedia.org/r/39044
[13:10:02] <gerrit-wm>	 Change merged: Hashar; [integration/jenkins-job-builder-config] (master) - https://gerrit.wikimedia.org/r/39045
[13:10:13] <gerrit-wm>	 Change merged: Hashar; [integration/zuul-config] (master) - https://gerrit.wikimedia.org/r/39046
[13:10:37] <gerrit-wm>	 Change merged: Hashar; [integration/zuul-config] (master) - https://gerrit.wikimedia.org/r/39047
[13:17:23] <gerrit-wm>	 New patchset: Hashar; "typo declaring the pep8 publishing macro" [integration/jenkins-job-builder-config] (master) - https://gerrit.wikimedia.org/r/39048
[13:17:40] <gerrit-wm>	 Change merged: Hashar; [integration/jenkins-job-builder-config] (master) - https://gerrit.wikimedia.org/r/39048
[14:12:00] <anomie>	 Nemo_bis, Nikerabbit - Why in I7fcbde6a did L10n-bot basically revert I50f775fd?
[14:22:04] <Nemo_bis>	 anomie: let's ask Raymond
[15:10:23] <Nemo_bis>	 anomie: Raymond_> Nemo_bis: it can happen when a i18n related patch set is merged in the middle of the export
[15:10:37] <Nemo_bis>	 he'll resubmit for you this evening before re-exporting
[15:12:39] <anomie>	 Nemo_bis- Thanks.
[17:40:52] <brion>	 Im guessing the wikimedia tech talk 12.13.12 is no longer "NOW"
[17:41:24] <^demon>	 Well, unless they've been running a talk for 4 days straight ;-)
[17:41:48] <brion>	 that's a long talk
[17:44:06] <^demon>	 brion: I've attended more than a few lectures that *felt* that long ;-)
[17:44:11] <brion>	 hehe
[17:46:13] <gerrit-wm>	 New patchset: Milimetric; "test" [analytics/reportcard/data] (master) - https://gerrit.wikimedia.org/r/39067
[18:37:08] <James_F>	 Krenair: Good luck with bug 14862 :-)
[18:37:53] <Krenair>	 James_F, thanks. csteipp's notes make this sound easier than I would've expected...
[18:38:22] <James_F>	 Krenair: Chris and I were talking about it quite a bit, and he got it all worked out, which I guess helps. :-)
[19:01:37] <preilly>	 ori-l: ping
[19:31:30] <hashar>	 hello
[19:32:43] <marktraceur>	 hashar: Hi!
[19:33:01] <^demon>	 hashar: Howdy. Did you see my e-mail?
[19:33:18] <hashar>	 ^demon: not yet
[19:33:40] <^demon>	 Subject: "Misconfigured Zuul?"
[19:33:48] <hashar>	 oops
[19:34:34] <hashar>	 so one job disappeared
[19:34:50] <^demon>	 That sounds...not good
[19:35:45] <Nemo_bis>	 unemployment can raise after elections
[19:35:46] <hashar>	 INFO:jenkins_jobs.builder:Creating jenkins job mediawiki-core-install-sqlite
[19:36:29] <ori-l>	 preilly: hey; in a meeting, but it's ending in 20. should i swing by yr desk?
[19:39:29] <hashar>	 ^demon: fixed
[19:40:28] <^demon>	 Cool, thanks!
[19:40:32] <^demon>	 Glad it was an easy fix :)
[19:41:20] <hashar>	 ^demon: also mailed wikitech about it as a post mortem
[19:44:34] <hashar>	 ^demon: can we get the Verified label extended tomorrow ? :)
[19:47:33] <^demon>	 Yeah, that should be fine.
[19:53:31] <StevenW>	 Do oldids change if you move a page?
[19:54:51] <Reedy>	 Can't see any reason why they would..
[20:03:46] <gerrit-wm>	 Change abandoned: Milimetric; "(no reason)" [analytics/reportcard/data] (master) - https://gerrit.wikimedia.org/r/39067
[20:57:09] <gerrit-wm>	 New patchset: Hashar; "support Verified+2" [integration/zuul-config] (master) - https://gerrit.wikimedia.org/r/39082
[20:57:50] <gerrit-wm>	 New review: Hashar; "Pending Gerrit change to extend Verified to -1..+2" [integration/zuul-config] (master); V: 0 C: -2;  - https://gerrit.wikimedia.org/r/39082
[21:10:05] <gerrit-wm>	 New patchset: Hashar; "patch merged is now pink" [integration/doc] (master) - https://gerrit.wikimedia.org/r/39083
[21:10:06] <gerrit-wm>	 New patchset: Hashar; "mention pipeline is from Zuul" [integration/doc] (master) - https://gerrit.wikimedia.org/r/39084
[21:10:06] <gerrit-wm>	 New patchset: Hashar; "update workflow with Verified+2 enhancement" [integration/doc] (master) - https://gerrit.wikimedia.org/r/39085
[21:10:34] <gerrit-wm>	 New review: Hashar; "Workflow flowchart is updated with https://gerrit.wikimedia.org/r/#/c/39085/" [integration/zuul-config] (master); V: 0 C: -2;  - https://gerrit.wikimedia.org/r/39082
[21:10:57] <gerrit-wm>	 New review: Hashar; "Wait for  https://gerrit.wikimedia.org/r/#/c/39082/ Zuul change to be applied" [integration/doc] (master); V: 0 C: -2;  - https://gerrit.wikimedia.org/r/39085
[21:11:18] <gerrit-wm>	 Change merged: Hashar; [integration/doc] (master) - https://gerrit.wikimedia.org/r/39083
[21:11:27] <gerrit-wm>	 Change merged: Hashar; [integration/doc] (master) - https://gerrit.wikimedia.org/r/39084
[21:15:04] <preilly>	 ori-l: I'm remote
[21:23:52] <ori-l>	 preilly: and how
[21:24:02] <preilly>	 ori-l: ?
[22:05:27] <Krinkle>	 hashar: So there's 2 things currently bothering me in ci that I'd like to know what your agenda is for.
[22:05:30] <Krinkle>	 * qunit
[22:05:57] <Krinkle>	 * build steps instead of job stacks
[22:05:57] <hashar>	 Krinkle: I am in conf call right now sorry
[22:44:16] <notpeter>	 fenari going down
[23:00:57] <^demon>	 notpeter: dive! dive!
[23:01:07] <notpeter>	 heh
[23:02:04] <hashar>	 anomie: so I will rebase the EQIAD related changes and ping you so we agree on deployment window
[23:02:37] <anomie>	 hashar- ok
[23:25:57] <Krinkle>	 James_F|Away: "Technical Debt" ?
[23:28:33] <Reedy>	 Krinkle: You know, there's this encyclopedia on the internet...
[23:28:33] <Reedy>	 http://en.wikipedia.org/wiki/Technical_debt
[23:30:09] <Krinkle>	 It is the name of a new component in VisualEditor bug tracker.
[23:32:11] <Krinkle>	 SauceLabs is awesome!
[23:32:33] <Krinkle>	 Reedy: Thx..
[23:39:07] <notpeter>	 running slightly over my window due to ops meeting/lunch combo
[23:47:33] <gerrit-wm>	 New patchset: Ori.livneh; "Subscribe to ZMQ publisher and log to time-rotated file" [analytics/glass] (master) - https://gerrit.wikimedia.org/r/39157
[23:48:47] <notpeter>	 fenari back up
[23:57:37] <gerrit-wm>	 Change merged: Ori.livneh; [analytics/glass] (master) - https://gerrit.wikimedia.org/r/39157