Time Nick Message 02:07 MTDiscord Agreed, hence why I'm putting forward a more flexible approach than hard coding in larger versions of biome mappings 10:31 sfan5 http://sprunge.us/vo5Gfe?diff pushing this soon 11:42 shaft LISTEN UP ENGINE DEVELOPERS. YOU FUCKED UP MT set_eye_offset and it behaves vastly different between 5.7/8 and the current dev build. We NEED your input to resolve how it should be handled for a mod and what is to be done about MT engine before 5.9 release. https://github.com/minetest-mods/ts_furniture/pull/20 11:53 [MTMatrix] I think you should insult us more, to obtain more authority and have us fix your issue in no time 11:54 shaft It's your fault so :) Should I put :) behind every sentence from now on? Just imagine screeming_anime_girl.png next to it with a memecenter label at the bottom right. 11:55 [MTMatrix] No, but you could fuck off with your manners, since we're all volunteering 11:55 shaft Just don't take everything so seriously. 11:55 MTDiscord Now that we have the protocol version bump in the release process, it should always be possible to detect which release somebody is running and work around any kind of behavior changes, even if they don't get caught or fixed in time for a release. 11:56 MTDiscord There is no need to panic. 11:56 [MTMatrix] warr1024: that doesn't seem fun though 11:56 rubenwardy blaming people doesn't result in a constructive environment. Bugs happen. The thing to do is provide reproduction steps, not point fingers 11:57 MTDiscord Lol, there is no need to NOT panic either. Do as you please 😄 11:57 shaft So I should put in a version check? Which function should I use? I know the behaviour started no longer being "broken" on 5.9 but when did it start? I know it's not working right in 5.8 and 5.7 but what about earlier versions? 11:58 MTDiscord You should file a bug report. Then politely escalate it if it seems major, or it's getting too close to a release and it's at risk of getting missed. 11:58 MTDiscord Then if it does get missed, the version check is a fallback plan. 11:58 shaft But I want the mod to be fixed now. 11:58 [MTMatrix] You should git bisect to find the commit at fault 11:58 shaft And I don't wanna debug the engine 11:59 [MTMatrix] I want to be rich, but people aren't going to give me money if I don't work 12:00 shaft I already made a PR and it works with the latest released version of minetest. I did work 12:00 [MTMatrix] Also, not even a company would fix a bug right away, so good luck 12:00 [MTMatrix] Sfan already told you why that PR is not a solution 12:01 shaft Can you please stop being passive aggressive toward me, Zughy? I said I'll put in version check. 12:06 [MTMatrix] You also said you want the bug to be fixed right away, said that your PR is fine when it's not and approached us pointing fingers. So, again, good luck. It's not being passive-aggressive. Personally, this conversation is over 12:07 shaft I said I want to fix the bug in the mod right away. 12:07 shaft And I will 12:09 MTDiscord MT does not support detecting specific non-release versions, like down to the commit granularity. If it hasn't been released yet then it's not a "version" of MT yet. 12:37 shaft I just detect whether it's 5.7.0 or 5.8.0. You never release new minor versions after all anway. 13:02 Desour LISTEN UP ENGINE DEVELOPERS and other people reading this. I FUCKED UP MT CMAKE_BUILD_TYPE, AND AS A RESULT TOLD YOU UNTRUE STUFF about RelWithDebInfo performance 13:02 Desour PR: #14600 13:02 ShadowBot https://github.com/minetest/minetest/issues/14600 -- Reject invalid CMAKE_BUILD_TYPE values by Desour 13:03 Desour it also explains why hotspot was suddenly no longer finding the symbols 13:05 Desour and I noticed, in irrlicht we don't have CMAKE_CXX_FLAGS_config for configs other than Release and Debug 13:05 Desour possibly also in other vendored libs, haven't checked 13:06 Desour so, irrlicht wont be properly optimized, nor have debug symbols, if built with RelWithDebInfo 13:06 json Hello! I'm using flatpak and I have a problem: even when I change the language in the settings, Minetest still uses English. Minetest version: 5.8.0; OS: Linux Opensuse leap 15.5. Thank you! 13:07 shaft I put version check in https://github.com/minetest-mods/ts_furniture/pull/20 but sfan5 got bored of testing it, so it would be appreciated if someone where to test and merge PR. 13:08 shaft *were to 13:17 json By the way, this issue has already been reported: https://github.com/minetest/minetest/issues/13210. Apparently this is a long-standing problem 13:18 json at least since version 5.6.1 13:18 shaft Use the Appimage https://github.com/An0n3m0us/Minetest-AppImages/releases 13:18 shaft Appimage rules 13:20 definitelya And goes back to talking to himself... 13:21 shaft You mean me? It was directed at Json who is using the flatpak and having trouble with the language settings. They work with the appimage 13:22 json OK 13:24 [MTMatrix] json: yeah, that's pretty annoying, I've been having the same issues for years 13:24 [MTMatrix] it either uses your system language or, if you change it, English 13:25 [MTMatrix] *same issue 13:25 json I'm just wondering: is this a Minetest or Flatpak problem? 13:25 [MTMatrix] also, meeting in about 1:30h 13:26 shaft It's a flatpak problem 13:26 shaft A minetest flatpak problem but such issues are typical for flatpaks 13:27 json Why? 13:30 shaft Because of the "isolation" they do. They do some half assed containering. It typically doesn't really add anything in terms of security. Flatpak just isn't made for users of home computers. 13:30 shaft Don't misunderstand: It works fine on home pcs but it isn't targeted at their use case 13:31 shaft It used to brake themes too. 13:32 json ok thank you 13:32 json good bye 15:00 [MTMatrix] dlin dlon, meeting time 15:01 [MTMatrix] pinging sfan5, srifqi 15:01 sfan5 I happen to be here 15:03 MTDiscord me too 15:03 [MTMatrix] Hoping more core devs will show up, we have just one point: "Feature freeze when? May 12, release June 1? (Zughy)" 15:04 [MTMatrix] (may 12 is when the next meeting is supposed to be) 15:04 Desour last release, the feature freeze far too long, imo 15:05 Desour it had to be enlonged because issues were not fixed in the expected time frame, and because some things breaking feature freeze had to be done 15:05 rubenwardy Freezes are supposed to only be a week 15:05 Desour so I would like it if we only enter feature freeze once the milestone is empty 15:05 [MTMatrix] That will probably never happen 15:07 pgimeno maybe it would be good to distinguish "milestone freeze" from "feature freeze"? 15:07 [MTMatrix] Also, all the issues in the current milestone are bugs, more specifically regressions https://github.com/minetest/minetest/milestone/23 15:07 [MTMatrix] and there are just two features (PRs) 15:10 sfan5 for me, feature freeze includes stopping accepting new features into milestone 15:10 sfan5 we could enter that phase any time but I fear working through the regressions is going to be a while 15:11 sfan5 (and to be honest I am not interested in working on the inventory regressions) 15:12 Desour just for reminder, the feature freeze took almost 2 months, from 2023-10-15 (according to meeting notes ) till around 2023-12-04 (release date according to blog) 15:12 [MTMatrix] I'm asking that for two reasons: 15:12 [MTMatrix] 1. I'd like to understand how feasible it is to feature the new content main menu section 15:12 [MTMatrix] 2. it helps us understand what to do 15:12 [MTMatrix] let's simply revert #13146 then (imagine a "juanchi face" emoji here) 15:12 ShadowBot https://github.com/minetest/minetest/issues/13146 -- Inventory mouse shortcut improvements by OgelGames 15:13 [MTMatrix] https://matrix.org/_matrix/media/v1/download/matrix.org/cQdiEaOxIAVBUHYOBtLcHKdB 15:13 [MTMatrix] and embrace the hate 15:13 [MTMatrix] (that was a joke, therefore the juanchi face 15:14 [MTMatrix] but it looks like the original author is not interested in fixing the regressions either :( 15:15 sfan5 if we find nobody to fix the regressions (or decide that they aren't so important after all) it may not turn out to be a joke 15:15 pgimeno I was thinking of feature freeze as: "no feature PRs are merged" which would imply "all feature PRs that are in the milestone are bumped to the next version" - so that new features have time to be tested (which is the point of a feature freeze) 15:16 [MTMatrix] I fear that this release doesn't offer much for modders 15:16 [MTMatrix] if we also remove that, which improves UX, meh.. 15:17 [MTMatrix] (looking at you, appguru, for #13987 👁️) 15:17 ShadowBot https://github.com/minetest/minetest/issues/13987 -- [pls do not squash :3] Allow limiting object observers by appgurueu 15:17 sfan5 I understand the marketing pov but the primary goal of a release is to get pending fixes and features into the hands of users not to create something exciting. 15:18 sfan5 see also: the "release early, release often" discussion 15:18 sfan5 anyway 15:18 [MTMatrix] for modders, there's the new and exciting async mapgen API 15:19 sfan5 my vote goes to we should enter soft feature freeze soon 15:19 sfan5 where soft feature freeze = no new features, except those in milestone (no new stuff added ofc) 15:20 Desour meh. features in the milestone are usually large features. blocking small features would be annoying 15:20 sfan5 well the point is to avoid new regressions during freeze 15:21 Desour then we should rather enter a freeze where we don't merge PRs that are likely to cause regressions 15:21 Desour (at least as long as they are not on the milestone) 15:23 sfan5 where's the difference to a feature freeze? 15:24 sfan5 I don't get your intent here 15:24 [MTMatrix] grorp: that's surely great news and it's been a huge amount of work, but not a lot of games use mapgen. On the contrary, something like entities invisible only for some players can be useful in a big plethora of genres; at least I think (in the end I don't use mapgen so I might not be impartial) 15:25 Desour in a feature freeze, we'd block *any* feature PR. this is different from blocking non-dangerous features 15:25 MTDiscord would gltf be considered admissible under the one approval rule? 15:26 sfan5 no 15:26 Desour luatic: model formats are part of the modding api, afaik 15:26 sfan5 a whole new model format is also anything except "relatively simple" 15:27 Desour the reason why we want to do feature freeze is so that we can be sure to test that the release version is bug free, right? 15:27 MTDiscord well it is unlikely to break anything except itself, assuming it is half-decently secure ;) 15:27 sfan5 yes 15:30 Desour so, if we do it earlier, it does not mean that we can release earlier. we still have to fix the bugs (which will definitely take more than a week). it doesn't matter if we fix them in a feature freeze or outside 15:30 sfan5 right 15:30 [MTMatrix] though a feature freeze may result in higher pressure to fix them fast 15:30 sfan5 I suppose the secondary function of a feature freeze is that it sets a "deadline" on when stuff should be finished so we can actually releasre 15:30 Desour therefore, I'd suggest we should find ways to make the bugs fixed faster, so that we can release earlier. and I don't think a freeze is the right tool for achieving this 15:31 MTDiscord we could just each pick a bug or two 15:31 [MTMatrix] Also, don't we incurr the risk to have too much on our plate in 5.10? From what I recall, there is the experimental GUI/HUD by v-rob, ambient light, various visual effects, CDB redesign, gltf format and possibly more. The risk of bugs seem pretty big, but delaying some of these features doesn't seem great either. Hence maybe diluting them (even just one) into 5.8 wouldn't be so bad 15:32 [MTMatrix] (*5.9 + grammar mistakes) 15:32 MTDiscord ambient light could make it into 5.9, it's relatively small and not that far from mergeability imo 15:34 MTDiscord it's something i'd very much like to still get into 5.9 for modders to play with in fact. i think it's relatively low risk, and has a good "bang for the buck" ratio. others and i have also already invested a good amount of time into reviewing and discussing it. 15:34 sfan5 just want to note that we can release as often and with whichever features we want 15:34 MTDiscord yes, but realistically, will we? 15:34 sfan5 the end result is not even much different if we were to release twice as often with half the features 15:35 sfan5 dunno 15:36 MTDiscord it's a discretization problem. i want to minimize the time modders have to wait for features like ambient light that are effectively "already there", hence if possible, i'd like to squeeze them in before the feature freeze. otherwise, they'll have to wait for another entire release period (which in principle could be arbitrarily short but practically will be a few months). 15:36 MTDiscord anyways, merging #14319 in 15m 15:36 ShadowBot https://github.com/minetest/minetest/issues/14319 -- Fix object:punch(puncher, ... should allow nil for puncher by sfence 15:39 Desour idea: to lure contributors into fixing milestone bugs, we could warrant them a special place in the credits as "release makers" or so 15:45 sfan5 do we have a consensus when to aim for a release? 15:46 sfan5 or just whenever it's ready? 15:47 [MTMatrix] And, do we wanna try shorter releases? 15:57 MTDiscord (start of) June seems reasonable, though i can't estimate how long it'll take to get rid of the regressions 15:58 MTDiscord an implicit shorter release "cycle" seems like a good idea 15:58 Desour +1 on both 15:59 sfan5 agreed 16:06 [MTMatrix] Nice, if there is nothing to add, meeting is over 16:26 [MTMatrix] Moved 5.9 milestone to June 2 (it's a Sunday), written down notes from today's meeting, created an internal discussion for the next one. Thank you all for participating 16:26 [MTMatrix] also reminder: vote for the new name 16:47 Niklp hi, https://github.com/minetest/minetest/commit/c524c52baa5456cff87127b3c4c41b5758b91c7a removed support for `set_physics_override(0,0,0)` which was never documented. I'm pretty sure it should be removed from the breakages.md file then (https://github.com/minetest/minetest/blob/72cb4e9bea803d6b300d095394e380c3674347c4/doc/breakages.md?plain=1#L14) 17:29 Krock Hi. will merge #14586 in 15 minutes 17:29 ShadowBot https://github.com/minetest/minetest/issues/14586 -- Client: fix unknown texture upon shift-move to full inventory list by SmallJoker 17:42 MTDiscord Was glTF discussed? 17:44 Krock Niklp: undocumented features should not need any special treatment. it's the modders fault for using it 17:44 Krock also: merging 17:44 Krock @josiah_wi Not in this meeting from what I can see. 17:47 MTDiscord realistically, it will probably have to wait till 5.10, since we want to release soon-ish, it's a rather large feature and there are still a good bunch of regressions to be fixed (and some smaller features that are farther along to potentially get in); it does not qualify for the new one approval rule 17:49 MTDiscord Yes, thank you. 21:51 MTDiscord If we can get through with #14543 into 5.9, I could help fixing regressions. 21:51 ShadowBot https://github.com/minetest/minetest/issues/14543 -- Add gameid aliases by nauta-turbidus