[04:26:17] <wikibugs>	 MediaWiki-extensions-MultimediaViewer, Analytics, Multimedia: Make upload.wikimedia.org serve images with Timing-Allow-Origin header - https://phabricator.wikimedia.org/T76020#939229 (Tgr)
[08:06:44] <wikibugs>	 Analytics-EventLogging, Continuous-Integration: UniversalLanguageSelector patches are blocked by voting schema tests - https://phabricator.wikimedia.org/T85124#939288 (Amire80) NEW
[08:24:04] <wikibugs>	 Analytics-EventLogging, ContentTranslation-Deployments, Continuous-Integration: UniversalLanguageSelector patches are blocked by voting schema tests - https://phabricator.wikimedia.org/T85124#939308 (Arrbee)
[08:54:46] <wikibugs>	 Analytics-EventLogging: EventLogging ValidateSchemaTest::testValidEvent() fails under HHVM - https://phabricator.wikimedia.org/T78680#939329 (QChris) Open>Resolved a:QChris >>! In T78680#932191, @hashar wrote: > Ideally we would have a test highlighting the issue.  If the issue was be the EventLoggin...
[09:58:07] <wikibugs>	 Analytics-EventLogging, ContentTranslation-Deployments: UniversalLanguageSelector patches are blocked by voting schema tests - https://phabricator.wikimedia.org/T85124#939427 (hashar)
[10:00:20] <wikibugs>	 Analytics-EventLogging, ContentTranslation-Deployments: UniversalLanguageSelector patches are blocked by voting schema tests - https://phabricator.wikimedia.org/T85124#939428 (hashar) The extension has bad resource loader definitions.  You can reproduce on your local setup using:   cd tests/phpunit  php phpun...
[11:55:35] <grrrit-wm1>	 (PS1) QChris: If desktop site starts in 'www.', drop 'www.' for mobile site [analytics/aggregator] - https://gerrit.wikimedia.org/r/181396
[12:16:14] <wikibugs>	 Analytics-EventLogging, ContentTranslation-Deployments, Continuous-Integration: UniversalLanguageSelector patches are blocked by voting schema tests - https://phabricator.wikimedia.org/T85124#939595 (hashar) So looking quickly at ULS, it depends on a RL module `schema.UniversalLanguageSelector` which is not i...
[12:19:59] <wikibugs>	 Analytics-EventLogging, ContentTranslation-Deployments, Continuous-Integration: UniversalLanguageSelector patches are blocked by voting schema tests - https://phabricator.wikimedia.org/T85124#939608 (Amire80) As I said, AFAIK the schema in ULS is loaded correctly using the hook, with the technique described h...
[12:29:11] <wikibugs>	 Analytics-EventLogging, ContentTranslation-Deployments, Continuous-Integration: UniversalLanguageSelector patches are blocked by voting schema tests - https://phabricator.wikimedia.org/T85124#939610 (hashar) The problem is the Jenkins job does not clone the EventLogging extension when preparing the tests for...
[12:48:47] <wikibugs>	 Phabricator, Analytics-Tech-community-metrics, Engineering-Community: Monthly report of total / active Phabricator users - https://phabricator.wikimedia.org/T1003#939636 (Aklapper) >>! In T1003#852987, @Dzahn wrote: > Access denied for user 'phstats'@...  [...] This requires a change in DB GRANTS.  @dzahn: Do...
[12:50:12] <wikibugs>	 Phabricator, Analytics-Tech-community-metrics, Engineering-Community: Monthly report of total / active Phabricator users - https://phabricator.wikimedia.org/T1003#939640 (Dzahn) >>! In T1003#939636, @Aklapper wrote: > @dzahn: Does that require a separate ticket, or how to proceed?  It does, but it already exi...
[12:53:46] <wikibugs>	 Multimedia, Analytics, MediaWiki-extensions-MultimediaViewer: Make upload.wikimedia.org serve images with Timing-Allow-Origin header - https://phabricator.wikimedia.org/T76020#939648 (Aklapper) Thank you that this is now a [[ http://www.google-melange.com/gci/task/view/google/gci2014/5832099869229056 | GCI ta...
[13:32:05] <wikibugs>	 ContentTranslation-Deployments, Analytics-EventLogging, Continuous-Integration: UniversalLanguageSelector patches are blocked by voting schema tests - https://phabricator.wikimedia.org/T85124#939691 (Amire80) Looks good to me now. Thank you.
[13:32:18] <wikibugs>	 ContentTranslation-Deployments, Analytics-EventLogging, Continuous-Integration: UniversalLanguageSelector patches are blocked by voting schema tests - https://phabricator.wikimedia.org/T85124#939692 (Amire80) a:hashar
[13:32:29] <wikibugs>	 ContentTranslation-Deployments, Analytics-EventLogging, Continuous-Integration: UniversalLanguageSelector patches are blocked by voting schema tests - https://phabricator.wikimedia.org/T85124#939693 (Amire80) Open>Resolved
[13:36:52] <wikibugs>	 operations, Analytics-EventLogging, Wikimedia-Site-requests: wikitech.wikimedia.org error  "$wgEventLoggingBaseUri is not set." - https://phabricator.wikimedia.org/T84965#939698 (Dzahn) a:Andrew
[15:07:25] <wikibugs>	 Analytics-Engineering, Analytics-Dashiki: Vital Signs user knows to click on metric title to open definition [3 pts] - https://phabricator.wikimedia.org/T76741#939856 (Milimetric) Open>Resolved
[17:25:59] <nuria>	  will be back in 30 mins
[17:43:41] <jdlrobson>	 YuviPanda: here :)
[17:45:04] <YuviPanda>	 jdlrobson: I poked ottomata in -operations too
[17:47:09] <jdlrobson>	 thanks YuviPanda heads up i have a standup in 15
[18:06:06] <qchris>	 milimetric: YuviPanda is asking me about the qchris-* labs instances. Will you still need them for the Storm experiments?
[18:06:10] <grrrit-wm1>	 (PS1) Mforns: Add Annotations API [analytics/dashiki] - https://gerrit.wikimedia.org/r/181424
[18:09:15] <milimetric>	 qchris: you are thinking of killing them entirely?
[18:09:25] <milimetric>	 i guess that's fine until I finish with the sandbox
[18:09:39] <milimetric>	 so far the sandbox has been enough
[18:09:51] <qchris>	 Yes, I am thinking of killing them entirely. It seems that would be preferable for YuviPanda.
[18:10:12] <milimetric>	 ok, that works for me qchris: it's not hard to bring them back right?  you have the puppetization saved?
[18:10:12] <YuviPanda>	 might be, at least partially because they’ll keep sending me 5 emails a day saying they’re down ;)
[18:10:41] <qchris>	 Ever-nagging Shinken :-P
[18:10:58] <qchris>	 milimetric: Yup, I have the puppet commits that I used to bootstrap them.
[18:14:47] <qchris>	 YuviPanda: Does Shinken allow you (non guest) to acknowledge the alerts?
[18:15:00] <YuviPanda>	 qchris: it does, but we don’t have a suitable user plugin yet
[18:15:53] <qchris>	 Sounds like I should ping you to acknowledge :-) Then at some point that'll get annoying for you, and booooom we have the user plugin? :-D
[18:16:00] <YuviPanda>	 :D
[18:16:05] <YuviPanda>	 just so manmy thing.
[18:16:15] <qchris>	 Ok. I'll kill the instances then.
[18:17:45] <qchris>	 Done.
[18:19:06] <grrrit-wm1>	 (PS2) Nuria: [WIP] Mobile apps oozie jobs [analytics/refinery] - https://gerrit.wikimedia.org/r/181017
[18:26:20] <YuviPanda>	 qchris_meeting: wheeee
[18:34:34] <grrrit-wm1>	 (CR) Ottomata: [WIP] UDF for classifying pageviews according to https://meta.wikimedia.org/wiki/Research:Page_view/Generalised_filters (4 comments) [analytics/refinery/source] - https://gerrit.wikimedia.org/r/180023 (owner: Ottomata)
[18:36:14] <Ironholds>	 ottomata, yay comments!
[18:38:06] <ottomata>	 just a few.  Ironholds i think there is one maybe that needs to be responded to still?
[18:38:15] <Ironholds>	 kk
[18:38:22] <ottomata>	 patchset 15 line 39
[18:38:28] <ottomata>	 about \\? in uriPath
[18:38:53] <jdlrobson>	 milimetric: you around?
[18:39:01] <milimetric>	 jdlrobson: yes
[18:39:01] <milimetric>	 hi
[18:39:28] <jdlrobson>	 so i was curious how i can test limn locally now.... i'm trying to work out how to make the urls in dashboards/report.json
[18:39:30] <jdlrobson>	 how do i do that now?
[18:40:23] <milimetric>	 jdlrobson: "make the urls"?
[18:40:32] <milimetric>	 like what they should be for a new graph that you've just added?
[18:40:49] <jdlrobson>	 milimetric: so currently the datasources are like http://datasets.wikimedia.org/limn-public-data/mobile/datafiles/edits-monthly-new-active.csv
[18:40:56] <jdlrobson>	 but i have a local file in datafiles/edits-monthly-new-active.csv
[18:40:58] <jdlrobson>	 that i want to use
[18:41:09] <milimetric>	 oh! ok
[18:41:15] <jdlrobson>	 http://localhost:5000/limn-public-data/mobile/datafiles/edits-monthly-new-active.csv doesnt seem to exist
[18:41:35] <milimetric>	 right, it's stupid... hang on
[18:41:42] <milimetric>	 i think i know but let me check
[18:43:18] <milimetric>	 jdlrobson: try: http://localhost:5000/data/datafiles/mobile/edits-monthly-new-active.csv
[18:43:24] <jdlrobson>	 yup that did it
[18:43:26] <milimetric>	 the two pieces of magic in there are:
[18:43:49] <milimetric>	 1. "mobile" - this is a keyword that limn uses to differentiate how it symlinks to your repository vs. other repos
[18:44:07] <milimetric>	 2. "data" in "data/datafiles" - this is just the route prefix that the limn server uses
[18:44:22] <milimetric>	 "magic" -> "trash"
[18:58:14] <grrrit-wm1>	 (PS1) Jdlrobson: Update scripts in light of recent changes [analytics/limn-mobile-data] - https://gerrit.wikimedia.org/r/181428
[18:58:15] <jdlrobson>	 ^ milimetric w00t!
[18:58:31] <jdlrobson>	 cc bmansurov:  ^
[19:00:08] <grrrit-wm1>	 (CR) Jdlrobson: [C: 2] Timebox before summing. [analytics/limn-mobile-data] - https://gerrit.wikimedia.org/r/181204 (owner: Bmansurov)
[19:00:15] <grrrit-wm1>	 (Merged) jenkins-bot: Timebox before summing. [analytics/limn-mobile-data] - https://gerrit.wikimedia.org/r/181204 (owner: Bmansurov)
[19:00:36] <milimetric>	 ha jdlrobson, ok, if that helps :)
[19:10:07] <tuxilina>	 hy, I want to log some events (these ones https://phabricator.wikimedia.org/T76416) and for that I am using EventLogging and EL dev server running on vagrant. But I can't seem to make this command 'eventlogging-devserver' to work. It gives me this output: http://pastebin.com/KmAwx1fA And if I try to run this command inside vagrant (from vagrant ssh) it says that I don't have permission (sudo chmod-ing is also not allowe
[19:10:07] <tuxilina>	 d). Any ideas on how to fix this?
[19:15:38] <nuria>	 tuxilina: hello, teh dev server has to be run "inside" vagrant
[19:16:10] <nuria>	 tuxilina: that is, you are instantiating that process inside vagrant's virtual machine
[19:17:08] <nuria>	 tuxilina: Did you see: http://www.mediawiki.org/wiki/Extension:EventLogging/Guide#Installing_the_EventLogging_devserver
[19:22:23] <tuxilina>	 nuria: yes, I followed those steps, but when trying to run 'python setup.py install' it gives me 'error: [Errno 13] Permission denied: 'eventlogging.egg-info/requires.txt'' even with root access.
[19:22:59] <nuria>	 tuxilina: Did you try running with sudo (while you are the vagrant user)
[19:23:26] <tuxilina>	 nuria: yes, the same
[19:25:09] <tuxilina>	 nuria: this is the output when run inside vagrant: http://pastebin.com/mJXNtNzA
[19:27:34] <tuxilina>	 I also asked on #wikimedia-dev earlier, but apparently it was a bit earlier and I thought that maybe here is best to ask. I'm sorry if I was wrong. Thing is I struggled with this for quite some time, and I tried to figure things by myself but at some point I am afraid of wasting too much time that can be solved quite easily with a question on IRC.
[19:29:03] <nuria>	 tuxilina: I would try uninstalling eventlogging.egg-info and trying again
[19:29:16] <nuria>	 tuxilina: when did you updated vagrant/ EL code  last?
[19:30:06] <tuxilina>	 yesterday I think, but not sure 100%.
[19:30:16] <nuria>	 tuxilina: i just tried the install with latest EL code and worked fine
[19:31:49] <nuria>	 tuxilina: http://stackoverflow.com/questions/1550226/python-setup-py-uninstall
[19:34:59] <nuria>	 ottomata: in order to make my coordinator run on a "daily" partition I need to define a datasets file like: https://github.com/wikimedia/analytics-refinery/blob/master/oozie/webrequest/datasets.xml
[19:35:09] <nuria>	 ottomata: without ${HOUR} correct?
[19:38:28] <ottomata>	 hm, you don't necessarily need a datasets.xml file, you can define the dataset within your coorindator
[19:38:36] <ottomata>	 we have a separate file so that we can include it in multiple coorindators
[19:39:19] <nuria>	 ottomata: right, but if i want to define "daily" datasets it seems a common case enough that should go in that file right?
[19:40:34] <ottomata>	 but, ja, it would be the same, except it would have a different frequency
[19:40:37] <ottomata>	 nuria: ja was about to say that
[19:40:51] <ottomata>	 i think we don't need an 'unchecked' daily dataset though...b
[19:40:53] <ottomata>	 but i'm not srue
[19:40:54] <ottomata>	 sure
[19:41:22] <nuria>	 ok, ottomata will make another dataset-daily.xml on that directory, is that correct?
[19:42:15] <ottomata>	 hm, no you can put it in the same file
[19:42:18] <ottomata>	 for daily webrequest
[19:42:23] <nuria>	 ottomata: ok
[19:42:24] <ottomata>	 just name it differently
[19:42:41] <ottomata>	 <dataset name="webrequest_bits_daily"> etc.
[19:42:42] <ottomata>	 i think is good
[19:42:56] <tuxilina>	 nuria: you said to uninstall but I cannot do that, it didn't install at all because it doesn't have permissions. And I tried after updating everything (vagrant and extensions).
[19:43:08] <nuria>	 ottomata: and the _SUCESS flag is also available at the daily level?
[19:43:17] <nuria>	 tuxilina: you can "remove' the egg directory
[19:43:41] <nuria>	 tuxilina: and start afresh perhaps?
[19:44:25] <ottomata>	 HMMMM
[19:44:27] <ottomata>	 wait a minute..
[19:44:32] <nuria>	 ottomata: so when partitions are processed we also setup a _SUCESS at: ${webrequest_data_directory}/webrequest_text/hourly/${YEAR}/${MONTH}/${DAY}
[19:44:37] <ottomata>	 lemme review some oozie docs, I might be steering you wrong
[19:45:02] <nuria>	 ottomata: maybe i can look this up directly in teh cluster actually, let me see
[19:45:52] <tuxilina>	 nuria: assuming the directory is in here "/usr/local/lib/python2.7/dist-packages/eventlogging-0.8_20141222-py2.7.egg", the dist-packages directory doesn't have any event logging egg file.
[19:47:29] <ottomata>	 no ok nuria
[19:47:34] <ottomata>	 i think you do not need a new dataset.
[19:47:39] <ottomata>	 instead, make your coordinator frequency be daily
[19:47:40] <ottomata>	 and
[19:47:51] <ottomata>	 HMM
[19:47:59] <ottomata>	 ah yes
[19:48:11] <ottomata>	 and then has <data-in> instances that specify a 24 hour period
[19:48:12] <ottomata>	 like
[19:48:28] <ottomata>	         <data-in name="input" dataset="logs">
[19:48:28] <ottomata>	           <start-instance>${coord:current(-23)}</start-instance>
[19:48:28] <ottomata>	           <end-instance>${coord:current(0)}</end-instance>
[19:48:28] <ottomata>	         </data-in>
[19:48:35] <ottomata>	 or the other way around even
[19:48:38] <ottomata>	 might be better?
[19:49:01] <nuria>	 tuxilina: look in /vagrant/mediawiki/extensions/EventLogging/server/eventlogging.egg-info
[19:49:49] <nuria>	 ottomata: I did that already, but in that case the coordinator is using also faulty partitions, is that not the case?
[19:50:22] <tuxilina>	 nuria: same, permission denied, even with sudo as a vagrant user, and even as root user.
[19:50:41] <ottomata>	 nuria, no, because the data-in event will not be available until all 24 datasets are ready
[19:50:50] <ottomata>	 hang on, linking to example...
[19:51:42] <ottomata>	 nuria: i'm have trouble linking
[19:51:49] <ottomata>	 go here:
[19:51:50] <ottomata>	 https://oozie.apache.org/docs/3.1.3-incubating/CoordinatorFunctionalSpec.html#a6.4._Asynchronous_Coordinator_Application_Definition
[19:51:56] <tuxilina>	 nuria: eventlogging.egg-info folder belongs to root:root . I don't know if that matters, while the others belong to user and group vagrant
[19:52:07] <ottomata>	 and search for the text "Coordinator application definition that creates a coordinator action once a day for a year, that is 365 coordinator actions"
[19:53:30] <ottomata>	 nuria, the idea is
[19:53:37] <ottomata>	 your coordinator's frequency is 1 day
[19:53:49] <ottomata>	 and it will use the previous 24 hourly instances
[19:53:58] <ottomata>	 as <data-in>
[19:54:19] <ottomata>	 in the coordinator then, you shoudl have reference to the coodinator's current time variables
[19:54:20] <nuria>	 tuxilina: likely you are having permits issues because your /vagrant directory (that is mounted to a dir to your local box) must have files you cannot delete from vagrant itself. I would start by removing eventlogging extension, making sure no code is left in vagrant or local box and doing the setup process within vagrant afresh. If you tried to 'install'
[19:54:20] <nuria>	 outside vagrant under the /vagrant mount, permits will conflict
[19:54:22] <ottomata>	 that you can pass to your workflow
[19:54:59] <nuria>	 ottomata: ok, but how does it to know to use the "cleaned" up partitions
[19:55:04] <ottomata>	 you'll probably need multiple <data-in> in <input-events>, one for each webrequest_source partition
[19:55:09] <ottomata>	 depends on which dataset you use
[19:55:15] <ottomata>	 if you use
[19:55:17] <ottomata>	 webrequest_bits
[19:55:22] <nuria>	 ottomata: ah , ok, i see
[19:55:35] <ottomata>	 the dataset will not be available until the _SUCCUSS flag exists
[19:55:36] <ottomata>	 if you choose
[19:55:41] <ottomata>	 webrequest_bits_unchecked
[19:55:48] <ottomata>	 it will be available as soon as the directory exists
[19:55:59] <tuxilina>	 nuria: thank you so much for your help. I am trying this now. Thanks!
[19:56:04] <ottomata>	 so, if you define a coordinator that spans 24 hourly datasets
[19:56:28] <ottomata>	 it will not be launched until each of the 24 directories has a _SUCCESS flag in them
[19:56:44] <nuria>	 ottomata: ok, understood now thank you
[19:56:49] <ottomata>	 yup :)
[20:12:03] <milimetric>	 ottomata: nice talk!
[20:12:10] <milimetric>	 at hakka
[20:12:51] <ottomata>	 thanks!
[20:14:05] <nuria>	 ottomata is a speaker ninja master, kia!
[20:28:33] <grrrit-wm1>	 (PS3) Nuria: [WIP] Mobile apps oozie jobs [analytics/refinery] - https://gerrit.wikimedia.org/r/181017
[20:48:17] <wikibugs>	 Analytics-Wikimetrics, Analytics-Dashiki: Vital Signs user sees annotations on graphs [13 pts] - https://phabricator.wikimedia.org/T78151#940453 (mforns) {F23564}
[20:49:30] <wikibugs>	 Analytics-Wikimetrics, Analytics-Dashiki: Vital Signs user sees annotations on graphs [13 pts] - https://phabricator.wikimedia.org/T78151#940454 (Milimetric) sweet :)
[20:49:31] <mforns>	 kevinator, yt?
[20:49:47] <kevinator>	 yeah im here...
[20:49:59] <kevinator>	 I’m in a cafe with a crappy connection, and it seems to be working.
[20:50:06] <kevinator>	 I was going to move in 10 mins
[20:50:19] <mforns>	 ok
[20:50:31] <mforns>	 do you want me to wait till youre back?
[20:52:22] <kevinator>	 nah, what’s up?
[20:52:37] <mforns>	 I uploaded a screenshot on Dashiki annotations in the task
[20:53:08] <mforns>	 I wanted to ask you for your opinion, and also ask if you think we should show it to Pau
[20:53:31] <mforns>	 maybe you or he want to change something
[20:53:42] <kevinator>	 ok, let me look at it
[20:53:50] <mforns>	 ok
[20:55:43] <kevinator>	 that’s cool.  can you show me the file the data is in?
[20:57:02] <mforns>	 kevinator, yes, just a sec
[20:57:09] <mforns>	 https://meta.wikimedia.org/wiki/Dashiki:RollingActiveEditorAnnotations
[20:58:30] <kevinator>	 mforns: do the annotations on the bottom appear in the order they are in the file?
[20:58:48] <mforns>	 no, they appear sorted by start date
[20:58:54] <mforns>	 the older first
[20:59:26] <kevinator>	 It’s weird that you can speficy an annotation for a date or range that is not even within the data range
[20:59:54] <kevinator>	 I think I should call a meeting with Pau to discuss next steps...
[21:00:13] <milimetric>	 kevinator: i would say that's not necessarily necessary at the moment
[21:00:15] <kevinator>	 otherwise this does seem useful when it will have real data in it
[21:00:31] <milimetric>	 we said when we tasked this that we'd have a second pass where we engaged Pau
[21:00:35] <mforns>	 exactly kevinator, Nuria, Dan and I were thinking in showing only the annotations that share the date range of the chart
[21:00:46] <milimetric>	 and that's when we would try to put the annotations in the graph itself
[21:01:23] <mforns>	 aha milimetric, so you think we should finish this task as it is now
[21:01:41] <milimetric>	 yes, that was how we envisioned it when we tasked it
[21:01:54] <mforns>	 and then create another task engaging Pau, ok
[21:02:03] <milimetric>	 yes
[21:02:05] <mforns>	 makes sense to me
[21:02:11] <kevinator>	 yes, a second pass later
[21:02:28] <milimetric>	 but this first pass was defined as: have annotations show below the graph with a little popup
[21:02:37] <milimetric>	 and it's great that you got markdown working, that's really neat :)
[21:02:43] <mforns>	 ok, so I'm going to finish it up
[21:02:45] <mforns>	 :]
[21:02:51] <mforns>	 that was easy, hehe
[21:03:03] <mforns>	 the layout was more tricky
[21:03:22] <kevinator>	 so is the width of the box determined by how much text is in it?
[21:03:36] <mforns>	 the enclosing box? no
[21:03:44] <kevinator>	 If we only have one annotation will it show up on the left? (left aligned?)
[21:03:46] <mforns>	 the note boxes, yes
[21:03:52] <mforns>	 yes, exactly
[21:04:57] <kevinator>	 one more question:  will each project:metric have its annotation file?
[21:05:15] <kevinator>	 It seems they may all inherit the same annotations from a metric
[21:07:05] <kevinator>	 This look great tho
[21:07:15] <mforns>	 kevinator, yes, now all projects share the same annotation file, and a metric will show all annotations regardless of which projects are being shown in the graph
[21:07:58] <kevinator>	 That’s fine for now.  We’ll see if we get any requests for project specific annotations
[21:08:09] <mforns>	 Dan and I spoke this the other day informally, maybe in the next step we can add 1-project-only support
[21:08:16] <mforns>	 yes
[21:08:45] <mforns>	 ok, so, let me finish this up and I let you guys know
[21:09:00] <kevinator>	 :-)
[21:09:14] <kevinator>	 i’m going to sign off for a bit and go back home
[21:09:19] <kevinator>	 I’ll be online again in 30 mins
[21:10:32] <mforns>	 ok
[21:33:04] <awight>	 Where should I request access to hive?  RT?
[21:43:49] <awight>	 Ironholds: I'm hoping to look up some personally identifying financial and health information in Hive.  I hear I can make queries from stat1002, but I don't see how to?
[21:44:08] <Ironholds>	 awight, type "hive" ;p
[21:44:19] <awight>	 baaa!
[21:44:22] <Ironholds>	 also, that's dumb
[21:44:26] <awight>	 you sneaky analsen
[21:44:31] <Ironholds>	 you know we keep the health information in MySQL
[21:44:36] <ottomata>	 haha, uhhh
[21:44:36] <Ironholds>	 hive is where we keep the social security numbers
[21:44:41] <awight>	 I just dump it everywhere
[21:44:50] <Ironholds>	 ahh, the AOL Approach
[21:44:52] <awight>	 can't be bothered to clean my dungeon
[21:44:55] <Ironholds>	 a valid analytics technique
[21:45:00] <ottomata>	 awight, you are not in the analytics-privatedata-users group
[21:45:05] * awight belches some bones
[21:45:07] <awight>	 o no!
[21:45:07] <ottomata>	 so you will not be able to access the webrequest data in hive.
[21:45:17] <ottomata>	 (if you can access that data, you had better let me know!)
[21:45:19] <awight>	 ottomata: is that an RT process to enlist myself?
[21:45:31] <ottomata>	 yes!  but RT is no more!
[21:45:36] <ottomata>	 Phabricator!
[21:45:55] <awight>	 ooo! thx
[21:46:02] <ottomata>	 the official access request policy in Phab is still being worked out.  but for now, if you create an issue and tag it with Operations or  maybe ops-request, i think someone will find it
[21:46:14] <awight>	 ottomata: confirmed that permission was denied :)
[21:47:08] <ottomata>	 good!
[22:14:15] <Ironholds>	 stupid C++
[22:14:33] <Ironholds>	 awight, you'll like this. One of my friends wrote a library for reading delimited files into R. It's 15x faster than the default.
[22:14:51] <Ironholds>	 one minor problem; somewhere in the ~1000 lines of C++, he accidentally, in the process of testing it, hard-coded it to only accept commas as separators
[22:15:03] <Ironholds>	 he can't actually work out where, we just know it blows up when you throw in a TSV.
[22:15:13] <awight>	 Ironholds: wow, I was totally trying to ignore your C++ emission above
[22:15:24] <Ironholds>	 Hey, I like C++. I just hate the people who write it.
[22:15:40] <awight>	 Bjarne is probably a really nice guy once you get past the whole C/C++ thing
[22:15:41] <Ironholds>	 It's like there's a Who Can Add Fewest Comments To Their Code competition and everyone upstream from me is invited.
[22:15:47] <awight>	 Ironholds: this code is open sourced somewhere?
[22:15:56] <Ironholds>	 oh yeah, https://github.com/hadley/fastread
[22:16:03] <awight>	 that is a heinous name
[22:16:10] <Ironholds>	 I found it in the end-  https://github.com/hadley/fastread/blob/88d4cfafbdcea174bd901438a7e9596de23f14c2/inst/include/fastread/SeparatorPolicy.h#L8
[22:16:23] <awight>	 also, problem with "read" is past tense vs imperative, etc...
[22:16:24] <Ironholds>	 the name is not the problem
[22:16:26] <awight>	 hah
[22:16:29] <awight>	 I can't get over it
[22:16:34] <Ironholds>	 the problem is that every R/C++ library I encounter is, like.
[22:16:39] <awight>	 yep
[22:16:40] <Ironholds>	 Like there's a guide to writing somewhere that says
[22:16:41] <awight>	 I remember
[22:16:43] <Ironholds>	 "do not add comments"
[22:16:47] <awight>	 oooh
[22:16:53] <Ironholds>	 "If you find you absolutely have to write comments, make sure they describe the code, not the why"
[22:16:53] <awight>	 and do evile magick if possible
[22:17:00] <awight>	 HAH
[22:17:02] <awight>	 thx for the link
[22:17:06] <Ironholds>	 so you end up with a thousand lines of C++ that include one comment which reads "set the namespace"
[22:17:13] <Ironholds>	 WHY. WHY ARE YOU SETTING THE NAMESPACE. TELL ME.
[22:17:19] <awight>	 not quite hardcoded but entertaining enuf
[22:17:32] <Ironholds>	 actually the library you read was a poor example. The code is mostly good. Just the lack of comments.
[22:17:38] <Ironholds>	 It balances out my C++ which is at least 30% comments.
[22:18:01] <awight>	 "Why am I doing this" :p
[22:18:10] <Ironholds>	 more https://github.com/Ironholds/urltools/blob/master/src/encoding.cpp
[22:18:35] <Ironholds>	 I have so much to make less-heinous in that library, but I am tremendously proud of the URL decoder/encoder.
[22:18:50] <awight>	 ohwow, "internal_url_encode"
[22:18:59] <Ironholds>	 Yes, I hate it too
[22:19:02] <awight>	 I like that your "%" fooled github
[22:19:10] <Ironholds>	 TL;DR Rcpp can't handle functions in classes
[22:19:17] <awight>	 WAT
[22:19:20] <Ironholds>	 if I pass encoder::url_encode() out to R, it starts screaming
[22:19:26] <Ironholds>	 "I DON'T KNOW WHAT THE ENCODER NAMESPACE IS HALP"
[22:19:38] <awight>	 oh.  mangling.
[22:19:52] <Ironholds>	 so you write things in classes, as god intended, and then https://github.com/Ironholds/urltools/blob/master/src/urltools.cpp#L56-L69
[22:20:22] * awight holds up gorgon shield
[22:20:26] <Ironholds>	 hahah
[22:20:35] <Ironholds>	 oh, you're lucky the guts are hidden for you.
[22:20:37] <Ironholds>	 *from
[22:20:37] <awight>	 Auugh I think it got me
[22:20:45] <Ironholds>	 if you want real fun, you should see R/C code
[22:20:49] <Ironholds>	 everything is a SEXP. EVERYTHING.
[22:21:02] <Ironholds>	 never before have so many caps locks sacrificed so much for so many
[22:21:09] <awight>	 please take out the MediaWiki style spaces around template brackets oh god
[22:21:27] <Ironholds>	 I like spacing it like that, it makes it readable
[22:21:44] <Ironholds>	 now if you insist on void func_name() { instead of void func_name(){ we will come to blows, good sir.
[22:21:47] <Ironholds>	 that is a stupid convention
[22:21:57] * awight abandons holy war
[22:22:02] <Ironholds>	 although less stupid than hurr durr let's put the brace on the next line hurr
[22:22:11] <Ironholds>	 that's just...why. why would you do that.
[22:22:13] <awight>	 I put them on a new line
[22:22:22] <Ironholds>	 why? Seriously, I'm interested to hear the rationale
[22:22:24] <awight>	 because then you can see the balanced pairs
[22:22:25] <Ironholds>	 I've never really got it
[22:22:31] <Ironholds>	 ...okay that's a really good point
[22:22:45] <awight>	 also, imagine a two-line else statement...
[22:22:59] <Ironholds>	 I just do if(foo){
[22:23:01] <Ironholds>	 bar;
[22:23:04] <Ironholds>	 } else {
[22:23:07] <Ironholds>	 baz;
[22:23:07] <Ironholds>	 }
[22:23:25] <awight>	 it's really easy and common to: else if (stupid condition)\n\t make my day;\n }
[22:23:37] <Ironholds>	 can we at least agree that the if(foo){bar} people need to go somewhere else?
[22:23:38] <Ironholds>	 hah. snap.
[22:23:50] <awight>	 anyway the brace on the same line is... K&R
[22:23:56] <awight>	 you want to be cool, don't you?
[22:24:12] <Ironholds>	 I want a lot of things but that actually isn't one of them.
[22:24:17] <Ironholds>	 Mostly I just want to be happy
[22:24:28] <awight>	 a list of string vectors.
[22:24:40] <awight>	 a single-quote before each line of documentation!
[22:24:44] <Ironholds>	 list < std::vector < std::string > >
[22:24:45] <awight>	 you're made of stouter stuff than me
[22:24:56] <Ironholds>	 oh, actually the single quote is for roxygen
[22:25:01] <awight>	 I'm sure...
[22:25:06] <Ironholds>	 it uses [comment_char]' in its parser
[22:25:12] <awight>	 you're not just rubbing that post-colonian shite in, you sure?
[22:25:13] <Ironholds>	 "this line has something to turn into a markdown page"
[22:25:22] <awight>	 *colonial*
[22:25:33] <Ironholds>	 what, with a single apostrophe?
[22:25:39] <awight>	 with half a quote
[22:25:45] <Ironholds>	 groooan
[22:25:57] <Ironholds>	 wait, awight
[22:26:11] * awight pretends was not wandering off already
[22:26:18] <Ironholds>	 would an apostrophe without a closing apostrophe be an unopposedtrophe?
[22:26:19] * awight satisfied to cause small cuts
[22:26:34] <awight>	 Ironholds: don't forget the newline and you shouldn't have that issue
[22:26:43] <Ironholds>	 fine, ignore my terrible pun
[22:27:08] <awight>	 it was fantastic
[22:27:30] <awight>	 It made me feel like the bar was closing.
[22:27:51] <Ironholds>	 I think that's a compliment and will interpret it as such. Thank you!
[22:34:55] <grrrit-wm1>	 (CR) Milimetric: [C: 2] Normalize project before deduplicating [analytics/wikimetrics] - https://gerrit.wikimedia.org/r/181001 (owner: Mforns)
[22:35:06] <grrrit-wm1>	 (Merged) jenkins-bot: Normalize project before deduplicating [analytics/wikimetrics] - https://gerrit.wikimedia.org/r/181001 (owner: Mforns)
[22:36:34] <wikibugs>	 Quarry: Make "Home" navlink go to profile for logged-in users. - https://phabricator.wikimedia.org/T85175#940705 (Capt_Swing) NEW a:yuvipanda