[08:04:31] * gehel is bringing Oscar to the eye doctor. Back as soon as I can
[10:36:55] <dcausse>	 ottomata: thanks for the task!
[10:36:57] <dcausse>	 lunch
[12:51:36] <inflatador>	 <o/
[12:52:32] <dcausse>	 o/
[13:42:03] <inflatador>	 dropping off kids, back in ~20
[14:08:42] <inflatador>	 back
[14:52:59] <ebernhardson>	 \o
[14:53:29] <dcausse>	 o/
[14:55:07] <ebernhardson>	 random thought, how are we going to auth flink w/ private wikis ?
[15:17:46] <dcausse>	 haven't about that :/
[15:38:12] <dcausse>	 wikiteck is not even using kafka, private wikis are not sending classic mediawiki events
[15:40:10] <ebernhardson>	 err...thats going to be difficult
[15:40:31] <ottomata>	 ^^ wut
[15:41:19] <dcausse>	 ottomata: this is from reading wgEnableEventBus 
[15:43:56] <ottomata>	 maybe we should change that?
[15:44:55] <dcausse>	 for wikitech even jobs over kafka were declined https://phabricator.wikimedia.org/T192361
[15:45:11] <ottomata>	 private doesn't use TYPE_PURGE either?   does that mean private wikitech pages caches are not purged?
[15:45:34] <dcausse>	 they're perhaps not even cached?
[15:45:38] <ottomata>	 i guess?
[15:47:07] <ottomata>	 maybe we can add TYPE_EVENT to all of those?
[15:47:09] <ottomata>	 i dont' see why not?
[15:48:36] <dcausse>	 I guess it depends on how HTTP public EventStream is going to not expose this
[15:48:46] <ottomata>	 ohhh right.
[15:48:47] <ottomata>	 right.
[15:48:51] <ottomata>	 hm.
[15:48:54] <ottomata>	 i do see why not.
[15:49:22] <dcausse>	 and then how enrichment is going to make mw api calls on those private wikis
[15:49:30] <ottomata>	 i think we'd need a way to change the stream name for those streams, and produce to different streams.
[15:49:38] <ottomata>	 oh api is behind auth? yeah
[15:49:40] <ottomata>	 nasty
[15:50:01] <dcausse>	 wondering how parsoid/restbase was dealing with those
[15:50:09] <ottomata>	 yeah indeed
[15:50:25] <ottomata>	 h they have TYPE_JOB though
[15:50:41] <ottomata>	 i think restbase is handled by jobs?  not sure
[15:50:46] <dcausse>	 calling RunJobs from changeprop might be special and might not require auth
[15:50:54] <ottomata>	 phewf
[15:50:56] <ottomata>	 nasty
[15:51:28] <dcausse>	 well I have no idea how this works, just guessing...
[15:55:02] <ottomata>	 you know, we have talked about emitting per wiki stream topics before, and then declaring an 'uber stream' that has all of the relevant topics in it.  if we did that, we could have a public one that doesn't incldue the private wiki topics
[15:59:16] <dcausse>	 yes... but realistically speaking I don't an easy way for us to support such wikis
[15:59:32] <dcausse>	 s/don't/don't see/
[17:29:57] <inflatador>	 lunch, back in time for pairing
[17:33:08] <dcausse>	 dinner
[18:07:29] <gehel>	 ryankemper, inflatador : too many meetings this evening. I'll skip the pairing session 
[18:17:29] <ryankemper>	 inflatador: gehel: ack!
[18:29:21] <inflatador>	 gehel np
[18:29:23] <inflatador>	 back
[18:30:39] <ebernhardson>	 meh okta always want a re-login :P
[19:39:10] <inflatador>	 workout, back in ~40
[20:36:15] <inflatador>	 sorry, been back
[20:55:28] * ebernhardson finally realizes .thenComparing has a second argument...trying to do Comparator.comparing(foo).reversed().thenComparing(bar).reversed() was doing my head in :P
[21:50:04] <ebernhardson>	 perhaps it's telling that someone created a website dedicated to picking the right way to put multiline strings in yaml :P https://yaml-multiline.info/