[10:56:55] <dcausse>	 lunch
[13:49:35] <gehel>	 I had a chat with Kirsten about T403593 / T403518. There is some urgency (end of October). I've added it to our current milestone.
[13:49:36] <stashbot>	 T403593: CirrusSearch should allow filtering on page creation and last edit timestamps - https://phabricator.wikimedia.org/T403593
[13:49:36] <stashbot>	 T403518: Add search filter for time since last edit - https://phabricator.wikimedia.org/T403518
[14:06:50] <ebernhardson>	 \i
[14:06:52] <ebernhardson>	 \o even
[14:31:05] <dcausse>	 o/
[15:53:50] <dcausse>	 hm.. cindy does not pull Depends-On on Elastica I thought it did
[15:54:35] <ebernhardson>	 dcausse: i think it only works for depends-on mediawiki-core
[15:54:44] <dcausse>	 ok
[15:54:53] <ebernhardson>	 dcausse: we basically take the hash and try and do a checkout in the root mediawiki dir, it doesn't know it's a different extension
[15:55:13] <dcausse>	 yes wondering if we can another try to Elastica?
[15:55:36] <ebernhardson>	 oh i guess we use `git review -d ...` and not checkout, but ya we could probably try elastica, or maybe we can ask gerrit what repo
[15:55:49] <dcausse>	 ok
[15:58:18] <ebernhardson>	 dcausse: apparently can get project with: https://gerrit.wikimedia.org/r/changes/?q=id=Ifaabc21838226a364ed0a2bb6473c00f9da3e8e4
[15:58:31] <dcausse>	 thanks
[15:58:38] <ebernhardson>	 but it's the annoying google response where you have to drop the first line because it's (intentionally) invalid json
[15:59:55] <dcausse>	 yes... that's what I'm seeing..
[16:01:13] <ebernhardson>	 i get that it's security...but it's just annoying everywhere.  i wonder how many problems are actually avoided with that
[16:06:23] <ebernhardson>	 oh apparently https://gerrit.wikimedia.org/r/changes/Ifaabc21838226a364ed0a2bb6473c00f9da3e8e4 is more direct (returns a single item instead of a query that returns an array of results)
[16:12:08] <dcausse>	 hm.. I wonder what it does when the change-id is backported to another branc
[16:12:10] <dcausse>	 h
[16:13:09] <ebernhardson>	 dcausse: lol, it responds with the plain text "Multiple changes found for I...."
[16:13:17] <dcausse>	 :)
[16:37:33] <dcausse>	 dinner
[16:56:31] <ebernhardson>	 meh, annoyingly QueryString class doesn't have the setMinimumShouldMatch
[17:06:17] * ebernhardson kinda wants an inheritance scheme for profiles...but not sure i want the complexity :P
[17:22:55] <ebernhardson>	 for some reason i thought we were still using FullTextQueryStringQueryBuilder in most places, but prod is set to perfield_builder everywhere which uses FullTextSimpleMatchQueryBuilder
[17:34:23] * ebernhardson also wants to somehow distinguish CirrusSearch\Query and CirrusSeach\Search.  They sound almost the same, once should be renamed :S
[17:35:35] <ebernhardson>	 i think it's something like query is for query building, and search is everything around query building (field definitions, result sets)
[17:42:05] <ebernhardson>	 i probably missed something, but it was indeed pretty easy to wire through the minimum_should_match.  does require a core patch and a new api param for the query-dependent profile
[18:45:02] <ebernhardson>	 huh, i didn't realize opensearch has range fields.  Like you can have a field that says a docs value is 12<x<16, then search for 15 and find it
[19:06:57] <inflatador>	 <o/
[21:02:20] <ryankemper>	 inflatador: 5m
[21:04:21] <inflatador>	 ryankemper ACK, I'm around
[21:10:20] <inflatador>	 ryankemper scratch that, I might have to meet with Comms about the Linux Fest presentation thing
[21:11:04] <inflatador>	 they pinged me about 10m ago, I'm gonna wait for them to respond...will join pairing once I'm done, assuming there's enough time left