[07:52:19] <dcausse>	 o/
[09:32:53] <pfischer>	 o/
[09:34:11] <pfischer>	 dcausse: You may have seen it already, but we do have a OS 3 test cluster on k8s since Friday (thanks inflatador!)
[09:38:06] <dcausse>	 pfischer: yes saw the message, will try to get some config done there soon
[10:25:17] <dcausse>	 errand+lunch
[14:13:43] <inflatador>	 <o/
[14:15:35] <dcausse>	 o/
[14:51:56] <inflatador>	 re: OpenSearch 3, I added T414697 to the Search board, if I can help LMK
[14:51:57] <stashbot>	 T414697: Build the required plugins for opensearch 3 - https://phabricator.wikimedia.org/T414697
[15:58:49] <dcausse>	 cleaning up glent tables manually to unblock the drop_data_daily dag
[16:00:16] <dcausse>	 inflatador: thanks, these plugins are built by upstream I'm not clear what we have to do on our side to get them deployed?
[18:35:08] <Trey314159>	 I figured out more about what's going on in T416764, and I thought it was interesting. IMEs generate additional javascript "composition" events and include the current word in the JS event. My guess is that the UI is sending that data instead of the full input to autocomplete. Also, you can try it on English Wikipedia with English words (and a Pinyin IME) and see what it does, so that's neat.
[18:35:08] <stashbot>	 T416764: Search box loads wrong suggestions when input CJK characters in multiple times - https://phabricator.wikimedia.org/T416764
[18:35:09] <Trey314159>	 I tagged some UI projects to hopefully get the right eyes on the problem.
[18:36:45] <dcausse>	 Trey314159: very nice, thanks!
[18:37:01] <dcausse>	 dinner