[09:56:10] <ejoseph>	 dcausse: i think i figured out the cirrusResultDump for the CompleteSuggester
[09:56:18] <dcausse>	 ejoseph: awesome!
[09:56:44] <ejoseph>	 if you have a few minutes you can help review it
[10:09:50] <ejoseph>	 errand
[11:35:08] <ejoseph>	 dcausse: I couldn't find the class for prefix search
[12:44:26] <dcausse>	 ejoseph: prefix search should be done via the Searcher class
[12:47:58] <dcausse>	 ejoseph: for reviews feel free to ship a patch in gerrit
[12:48:13] <dcausse>	 even if it's not totally finished
[12:52:05] <dcausse>	 hm... but what you've done already might just be enough, you can test using prefixSearch by forcing the "classic" profile 
[12:52:12] <dcausse>	 e.g. /w/api.php?action=opensearch&search=main&profile=classic&cirrusDumpQuery
[13:05:42] <inflatador>	 Greetings
[14:20:37] <cbogen_>	 hey folks, especially ebernhardson - I want to check on T304954 and T305851, I know they were blocked by some SD work but my understanding is they are unblocked now. Can they be picked up? Do you need anything else from SD on this? Reminder that we need them done by the end of the month.
[14:20:38] <stashbot>	 T305851: Import has-suggestions flags to search indices - https://phabricator.wikimedia.org/T305851
[14:20:39] <stashbot>	 T304954: Import data from hdfs to commonswiki_file - https://phabricator.wikimedia.org/T304954
[14:27:58] <dcausse>	 cbogen_: I think Erik was waiting for some hive tables but I don't see any new comments related to this I'll ask Cormac
[14:28:09] <cbogen_>	 the task that made the new hive tables was resolved
[14:28:23] <cbogen_>	 T307983
[14:28:24] <stashbot>	 T307983: Write search index data for image suggestions into a hive table rather than local hdfs files - https://phabricator.wikimedia.org/T307983
[14:53:47] <dcausse>	 errand
[15:01:40] <ebernhardson>	 \o
[15:23:56] <dcausse>	 o/
[15:28:55] <ebernhardson>	 cbogen_: i'm a bit confused why the same table is mentioned in both tickets, are these tickets for the same data?
[15:31:43] <inflatador>	 will not make Unmeeting today, still trying to clean up stuff before end of week
[16:00:11] <dcausse>	 what room are we?
[16:01:39] <ryankemper>	 dcausse: it says community building room
[16:05:37] <cbogen_>	 ebernhardson: good question, I think it's different data but coming from the same table, but Marco or Cormac would be better to confirm - can you get in touch with them?
[16:28:36] <inflatador>	 workout, back in ~30
[16:45:02] <ryankemper>	 Anyone know the answer to the question in https://phabricator.wikimedia.org/T306899#7943304 as far as why we closed https://phabricator.wikimedia.org/T306919 as a duplicate? I think I remember us saying that the problem was the errors themselves and not the error message thus why it was marked duplicate but my memory's a bit tenuous
[16:48:23] <ebernhardson>	 while not why we closed it, i guess i don't know what a better 500 message is. When i hear that i think fail-whale, but I wouldn't really consider that better
[16:48:52] <inflatador>	 back
[16:50:53] <ryankemper>	 ebernhardson: yeah agreed, it seems not a good idea for us to make the error message assume 500s are always an issue on our end, even though in practice right now that's probably true 99.9% of the time
[16:51:14] <ryankemper>	 The real problem is more the existence of the errors than the message itself
[16:51:59] <cbogen_>	 My recollection is we closed it because we made improving the error message part of the acceptance criteria for T306899
[16:52:00] <stashbot>	 T306899: WCQS 500 errors - https://phabricator.wikimedia.org/T306899
[16:52:22] <cbogen_>	 The AC now says "the error message is understandable by users
[16:53:39] <ryankemper>	 cbogen_: Ah yes that matches up with my memory! Thanks, I'll respond with that
[16:54:07] <cbogen_>	 ryankemper: thanks!
[18:05:23] <inflatador>	 lunch, back in ~30-45
[19:17:03] <inflatador>	 back