Time Nick Message 11:46 rubenwardy oh great, GitHub artifacts aren't public 12:08 sfan5 yea 12:08 sfan5 gitlab's are though 12:09 Krock i.e. https://gitlab.com/minetest/minetest/pipelines/ 12:54 MTDiscord What? Github artifacts aren't public, wtf. That's the whole point. Now wonder everyone likes gitlab 13:10 Zughy[m]1 Please remind me: what are the problems with switching to GitLab? 13:11 rubenwardy GitHub fanboys don't like it 13:11 rubenwardy Jokes, I think the main concern is effort 13:12 rubenwardy GitLab is also slow, the UI is complicated, and it doesn't support Windows VM builds for pipelines 13:12 Zughy[m]1 can issues and PRs be "ported"? 13:12 rubenwardy yeah, that's an automatic importer 13:12 rubenwardy it breaks authors and subscriptions though, ofc 13:13 rubenwardy the issue for this was https://github.com/minetest/minetest/issues/7412 13:13 sfan5 https://github.com/minetest/minetest/issues/7412#issuecomment-397821719 13:14 Zughy[m]1 I'm too lazy to thumb down that comment 13:14 MTDiscord Having both is a good strategy 13:15 Zughy[m]1 about UI: I started with GitLab and for me it's quite the opposite. I guess it's relative, it depends on your first git site 13:15 rubenwardy An objective reason is to be more accessible to those who don't want to use a propietary platform, which is what GitHub is. There have been numerous contributors who have been unable to contribute because of this 13:15 rubenwardy pgi*meno being one example 13:16 rubenwardy I have an issue as an alternative to this: #7901 13:16 ShadowBot https://github.com/minetest/minetest/issues/7901 -- Accepting PRs/patches from non-Github users 13:16 Zughy[m]1 Yeah, I basically forced myself to register to GitHub just because of Minetest. It's literally the only GH project I contribute to, but I'm not happy (at all) to do my things on a Microsoft owned company 13:20 Zughy[m]1 huh, that's the issue where Paramat implicitly called me an extremist, how nice of him 13:23 rubenwardy accepting MRs from GitLab is probably the best compromise we'll do 13:25 Zughy[m]1 Out of curiosity: how many core devs, if it were up to them and them only, would like to switch? 13:25 rubenwardy you could move the issues to something like bugzilla, hopefully that has better issue management tools. I find it quite painful on GH 13:25 rubenwardy Eh. I've moved all my personal projects to GitLab, but I don't have as many issues or PRs 13:25 rubenwardy Breaking the author infomation for all the issues and PRs is quite annoying 13:27 celeron55 i still think github's status as the de facto git platform benefits MT 13:27 celeron55 it's why MT moved on github, and it seems to end up being the last reason to be there 13:27 celeron55 onto* 13:40 kilbith there should be a trusted member who decides who's allowed to translate on weblate, for each language 13:40 kilbith because I was shocked by some french sentences in the git log 13:43 Krock do we need a formspec version bump? 13:43 kilbith 5.4 has bumped to 4 already 13:44 kilbith (iirc) 13:44 rubenwardy It looks like weblate also has lots of incorrect translations 13:44 rubenwardy like, it's set translations for "Base game:" to translations for other things 13:44 rubenwardy such as host a game 13:44 kilbith rubenwardy, it's just completely abnormal that random people are allowed to translate liberally 13:45 Krock ah yes. 5.3.0 was formspec version 3. should be fine. 13:45 rubenwardy well, weblate doesn't even let you ban people. So I definitely support changing it to invite only 13:46 Krock kilbith: for comparison, how other games do it: https://twitter.com/Dwimepon/status/1355464537093373952/ 13:46 kilbith like some Japanese could insert "dick" in the translations without anyone noticing it 13:48 rubenwardy Wuzzy could be the person to do this 13:48 rubenwardy also - there appears to be no way to list recent changes either 13:48 kilbith one trusted member per language, possible? 13:49 sfan5 rubenwardy: those are fuzzy translations automatically determined to possibly help translating stuff 13:49 rubenwardy oh right, are they not included in .po? 13:49 sfan5 they are but they won't be shown ingame 13:49 rubenwardy that's good 13:50 sfan5 rubenwardy: there is a way to list recent changes and no, making a single person the gatekeeper to all translations won't scale 13:51 sfan5 the easiest way to check translations is in fact to look at the commits 13:51 sfan5 I give some a cursory look before pushing the translations 13:51 rubenwardy well, I mentioned recent changes because you can use that to spot bad translations. But commits do do that 13:52 rubenwardy and definitely don't require a single gatekeeper, but do have translators whitelisted maybe - given that banning isn't possible 13:54 sfan5 I don't know if kilbith was referring to bad translations or actualy vandalism, but for translations where we have some dedicated people that take care of it it'd be a good to have it be whitelist/invite-only 13:54 sfan5 but for languages nobody speaks not so much, it'll only hinder progress 13:54 rubenwardy you can only set it for the entire project, it's not perlanguage 13:55 rubenwardy https://hosted.weblate.org/settings/minetest/#access 13:55 kilbith no just bad translations regarding french commits 13:56 kilbith weblate may have a way to detect vandalism tho 14:07 Krock sfan5: what's removed first, the player from ServerEnv, or the PeerID? 14:08 Krock https://github.com/minetest/minetest/pull/10887/files#diff-97dfad7082316c04d6b4e4fd3fa0235c30464bfa799ebf5950c7fadb9d5cdf57R123 14:08 Krock this first checks against the RemotePlayer in ServerEnv, but latter code requires RemoteClient 14:08 Krock so if those two are unloaded independently there's still a race condition issue 14:10 Krock okay. according to PlayerSAO::unlinkPlayerSessionAndSave PeerID and RemotePlayer are removed at the "same" time 14:11 Krock so RemotePlayer is removed first, RemoteClient afterwards 14:16 Krock s/is removed first/is marked for removal first/ 14:26 sfan5 Krock: the peer can disappear from the connection at any time, sometimes after that happens the client+player is actually removed from the server 14:26 sfan5 s/sometimes/some time/ 14:27 Krock but RemotePlayer is removed by the SAO unloader function after it's marked for removal 14:27 Krock and RemoteClient is removed right away after on_leaveplayer is executed 14:28 sfan5 so Client and Player are removed at the same time, right? 14:32 Krock Client is removed in Server::DeleteClient -> handlePeerChanges -> AyncRunStep. Player is removed in PlayerSAO::removingFromEnvironment() called from ServerEnvironment::removeRemovedObjects() called from ServerEnvironment::step() 14:32 Krock AsyncRunStep vs step 14:32 sfan5 AsyncRunStep calls step itself 14:33 Krock ah I see. so client first, player afterwards in the same AsyncRunStep 14:33 Krock kinda complicated but OK since Lua is running in the same thread anyway 14:35 sfan5 From what I understand the assumption that {RemoteClient, RemotePlayer} will always exists together is safe 14:35 sfan5 in addition to that when PlayerSAO exists {RemoteClient, RemotePlayer} will also exist 14:36 Krock yes but only because both, the peer disconnect and entity deletion are handled in the same step 14:37 Krock anyway, I cannot find a way to trigger the (previously) buggy conditions for #10887 but since this is discussed I don't have any concerns 14:37 ShadowBot https://github.com/minetest/minetest/issues/10887 -- Cache client IP in RemoteClient so it can always be retrieved by sfan5 14:37 Krock looks okay, and the API still produces the same results 17:06 rubenwardy bug fix for a bug from 2014: #10890 17:06 ShadowBot https://github.com/minetest/minetest/issues/10890 -- Fix minetest.dig_node returning true when node isn't diggable by rubenwardy 17:12 sfan5 sounds good but could this break if mods don't return true? 17:14 rubenwardy nil is interpretted as the same as true 17:14 rubenwardy so if mods just return nothing, then it'll be the same as before 17:14 rubenwardy > bool result = lua_isnil(L, -1) || lua_toboolean(L, -1); 17:16 rubenwardy hmm, mods may return nil expecting it to mean false 17:17 rubenwardy I could add a deprecation 17:26 sfan5 nah that's fine 17:26 sfan5 please add a comment to the C++ method though 17:27 rubenwardy done 17:30 rubenwardy are your utf16 changes likely to fix #2574 ? 17:30 ShadowBot https://github.com/minetest/minetest/issues/2574 -- Character conversion failed 18:07 sfan5 I don't know, I planned to ask in that issue after it was merged 18:32 rubenwardy merging #10890 in 10 18:32 ShadowBot https://github.com/minetest/minetest/issues/10890 -- Fix minetest.dig_node returning true when node isn't diggable by rubenwardy 18:37 rubenwardy #3105 18:37 ShadowBot https://github.com/minetest/minetest/issues/3105 -- fgettext issues 18:37 rubenwardy >2015 18:37 rubenwardy >trivial 18:39 MTDiscord I guess #9315 was left to collect dust for another few months :/ 18:39 ShadowBot https://github.com/minetest/minetest/issues/9315 -- Require debug priv to view gameplay-relevant debug info (2nd try), add wireframe priv by Wuzzy2 19:26 sfan5 merging #10887, #10885 in 10m 19:26 ShadowBot https://github.com/minetest/minetest/issues/10887 -- Cache client IP in RemoteClient so it can always be retrieved by sfan5 19:26 ShadowBot https://github.com/minetest/minetest/issues/10885 -- Better enable_damage=false / 'immortal' group fix by sfan5 22:06 sfan5 rubenwardy: is there a planned fix for #10657 22:06 ShadowBot https://github.com/minetest/minetest/issues/10657 -- Unable to download mods from browse online content (Download Denied, Temp folder isn't a allowed path) 22:06 sfan5 ? 22:14 rubenwardy I don't have Windows 22:17 rubenwardy I could mess with the tempdirs but I wouldn't be able to test it 22:18 rubenwardy Perhaps the fix is to change os.tempfolder() to use GetTempPath on windows 22:18 rubenwardy rather than os.getenv("TEMP") 22:23 rubenwardy Well, or just use completely the same implementation 22:31 rubenwardy We should just close PRs which don't follow the PR template, like #10798 22:31 ShadowBot https://github.com/minetest/minetest/issues/10798 -- Fix handling of empty nodemeta by appgurueu 22:35 sfan5 hm dunno 22:35 rubenwardy The reason that the how to test section exists is to make sure that authors have thought about how their PR can be tested 22:36 sfan5 regard that I implicitly read the lack of How to test section as a "see linked issue" 22:36 sfan5 regarding* 22:37 rubenwardy I guess