Time Nick Message 00:01 paramat good 00:35 sofar sfan5: https://i.imgur.com/87LrMhL.png - something as simple as this could be a good default for latency and thus ordering 05:34 ANAND #8500 05:34 ShadowBot https://github.com/minetest/minetest/issues/8500 -- util/travis/: Allow common.sh and lint.sh to be executable by ClobberXD 09:04 sfan5 sofar: i'd maybe s/Shanghai/Hong Kong/, looks pretty good otherwise 10:40 ANAND Does our LINT work by simply calling clang-format without the -i switch? 10:41 ANAND Hmm... that doesn't produce the diff like our LINT does 10:42 ANAND It'd be nice if the LINTER's output can be directly applied as a patch :) 10:44 ANAND Ah I get it now. lint.sh passes the clang-format output to diff for every file 10:56 ANAND I suppose I could use clang-format directly 11:05 ANAND #8502 11:05 ShadowBot https://github.com/minetest/minetest/issues/8502 -- Fix code-style in src/client/ by ClobberXD 11:12 sfan5 this is likely to be rejected for the same reason the other PR was 11:13 sfan5 merging #8494 in 5 minutes 11:13 ShadowBot https://github.com/minetest/minetest/issues/8494 -- [NO SQUASH PLS] PlayerSAO::setHP - Don't call on_hpchange callbacks if HP hasn't changed by ClobberXD 11:31 ANAND sfan5: Likely... What's the better approach here? 10 files per PR? 11:31 sfan5 don't fix it at all 11:31 ANAND Heh 11:34 ANAND How about fixing code-style of only the touched files of the PRs I make? 11:35 sfan5 that makes them hard to review 11:35 ANAND :( 11:36 ANAND Hard to review if I squash all the changes into a single commit. What if the code-style commit is separate from the PR's main commit? 12:07 ANAND sfan5: #8494 :P 12:07 ShadowBot https://github.com/minetest/minetest/issues/8494 -- [NO SQUASH PLS] PlayerSAO::setHP - Don't call on_hpchange callbacks if HP hasn't changed by ClobberXD 12:08 sfan5 oh i forgot again 13:16 ANAND :) 15:52 sofar sfan5: maybe kinshasa or lagos for africa, johannisburg is really far south 15:53 sofar sfan5: but yeah, I think a simple middle point latency is a good start 21:43 sfan5 sofar: by the way, returning continent with list.json is a problem because list fetches right now are just disk reads 21:44 sfan5 uwsgi is not even involved 21:44 sofar well 21:44 sofar you only have what, 7 continents? 21:44 sofar could make 7 static files, one for each continent 21:44 sofar but yes, you'd need to geoip every incoming request 21:44 sfan5 surely we can spare a second http endpoint 21:45 sofar oh, yes, absolutely 21:45 sofar "tell me my region" 21:45 sfan5 it's 6 btw, NA, SA, EU, AF, AS, OC 21:46 sofar and have the client cache it for at least the session 21:46 sfan5 yeah 21:46 sofar hey, someone might play on antarctica 21:46 sofar I didn't even think of Oceania 21:47 sfan5 I thought that's why you included Auckland 21:47 sofar australia 21:47 sofar I don't know what continents the kids learn nowadays 21:47 sfan5 Auckland is in New Zealand though... 21:48 sfan5 if you ask me Oceania is AU, NZ and all those tiny islands around there 21:48 sofar oh, ha, NZ is not on the australian continent 21:48 sofar shame on me 21:49 sofar https://en.wikipedia.org/wiki/Oceania 21:49 sofar looks like for OC we should just pick sidney given that it's both geographically central and the largest population center of it 21:50 sfan5 probably 21:50 sofar so you would be OK with a simple uwsgi job that geolocates for clients? 21:51 sfan5 yea 21:51 sofar ok, then that's a fun weekend project we can toy around with 21:52 sfan5 -rw-r--r-- 1 root root 1436463 Mar 15 2018 /usr/share/GeoIP/GeoIP.dat 21:52 sfan5 not directly related, but I wonder if this is Ubuntu's fault 21:56 sfan5 ubuntu does not package the non-legacy variant of MaxMind's database, cool 21:57 sofar well whatever is your preference for running it, I'd say 21:57 sofar I'd code it in Golang myself tbh, but if you're already using uwsgi 21:57 sofar I just don't fancy python that much :)