Time Nick Message 09:09 sfan5 merging #13284 and #13627 09:09 ShadowBot https://github.com/minetest/minetest/issues/13284 -- Try all known video drivers if the requested one fails to initialize by numberZero 09:09 ShadowBot https://github.com/minetest/minetest/issues/13627 -- Allow running individual tests by numberZero 15:27 sfan5 range_all=1 does not turn off fog for clouds, is that expected? https://0x0.st/HQju.png 15:27 sfan5 cc @x2048 15:30 Krock looks like it's not expected 15:35 Zughy[m] meeting in about 2:30h 15:46 MTDiscord @x2048 15:48 sfan5 when generating entirely new blocks the engine seems to create blank blocks filled with ignore and then spend time copying the contents into the emerge thread, doesn't seem ideal 15:54 rubenwardy Is that to persist overgenerated features? 15:55 rubenwardy Like when trees/caves leak into ungenerated mapblocks 15:55 sfan5 yes 15:55 sfan5 it's just that it shouldn't spend time on blank blocks 15:56 sfan5 MapgenSinglenode also spends lots of time in updateLiquid even though it only ever generates air 16:01 sfan5 plus updateLiquid isn't really all that efficient with constant nodedef lookups 16:04 Krock could be made more efficient by caching liquid content IDs if that does not already happen 16:07 sfan5 it doesn't but there's also something simpler that should help a lot 16:54 sfan5 i have a singlenode map and faster flying speeds, if I fly forward I often hit ignore walls until they generate 16:54 sfan5 perf shows the emergethread not being busy all the time 16:54 sfan5 I wonder what's limiting thing 16:54 sfan5 this* 16:59 sfan5 curiously this only happens above ground, not on ground (if I enable a flat mapgen) 17:00 sfan5 different thought: I think we should have a chatcommand that allows a client to view the server profile results in realtime like F6 does on the client. profiler_print_interval is too annoying to work with. 17:13 sfan5 re: limiting thing it's not the emerge queue and it's not max_simultaneous_block_sends_per_client 17:14 sfan5 the issue disappears if you don't look forward when flying, so some of the estimation logic on the server is broken 17:37 sfan5 // HOTFIX: [...] Somebody have to find out whats wrong with the math here. 17:37 sfan5 you love to see it 17:50 Krock TODO FIXME 17:51 Krock > the issue disappears if you don't look forward when flying 17:51 Krock pretty sure a sign error then 18:05 Krock okay. it's time for a short meeting. 18:06 sfan5 I wrote a simple unit tests and it doesn't fail terrible, so it's not like the sign is just flipped 18:06 Pexin does it behave the same when flying backwards in ANY direction or just certain directions 18:06 Krock https://dev.minetest.net/Meetings#2023-06-25 18:08 sfan5 hmm 18:10 sfan5 > Start considering a hypothetical new name, possibly privately. See Here (Zughy) 18:10 sfan5 I support this and it should happen privately at first 18:11 Pexin wait so is zug's iron fist thread deleted? I haven't kept up 18:12 Pexin (nm I'll check) 18:12 Krock there are many suggestions but the most popular would need another poll iteration to narrow down the options 18:12 nrz_ hello, oh i'm on time for a meeting ? 18:12 Zughy[m] I'd be really glad if people stopped addressing a few rules to keep the topic in order as "iron fist" :\ 18:12 nrz_ name, you mean project renaming ? 18:12 Krock for me this does not have much priority because Minetest is a fine name IMO 18:13 nrz_ it's also our known name for ages 18:13 Zughy[m] I basically agree with sfan5 18:13 nrz_ if we want to rename, maybe provide 3 name and make it voted by community ? 18:14 Zughy[m] if you do, some lamers will buy the domain and we'll be stuck with the same problem we currently have with minetest.net/org 18:14 Zughy[m] (I didn't want to send a link, I meant .net and .org) 18:15 nrz_ yeah right, we already discussed this topic some time ago 18:15 sfan5 it should be considered at least, no guarantee that we can decide on a new name 18:16 sfan5 because e.g. I can understand how the name can be misleading for new people 18:16 Zughy[m] I've been to 5 different conferences with Minetest in June and yes, it's very misleading 18:20 Krock sfan5: for the "isBlockInSight" issue.. it might be possible to recycle #13615 with its cylindrical approximation, convert it to sphere and call it a day 18:20 ShadowBot https://github.com/minetest/minetest/issues/13615 -- ActiveObjectMgr::getActiveSelectableObjects by numberZero 18:21 sfan5 dunno 18:21 sfan5 I'll write some unit tests first, see if they fail and then consider rewriting the function 18:21 Krock seems at least to be a more geometric solution than the cosine and dot product war that's going on in this function 18:22 Krock alright 18:22 Zughy[m] moving on 18:22 Zughy[m] "Less than two months to 5.8: we should consider what to do with a few issues/PRs in the milestone to avoid a last month burst (Zughy)" 18:22 Zughy[m] milestone for you: https://github.com/minetest/minetest/milestone/22 18:23 Zughy[m] My concerns are mostly towards #11016 and the big list in #13476 18:23 ShadowBot https://github.com/minetest/minetest/issues/11016 -- Dual Wielding by LizzyFleckenstein03 18:23 ShadowBot https://github.com/minetest/minetest/issues/13476 -- Settings UI mega-issue 18:23 sfan5 my contribution 🤷 18:24 Krock the milestone issues and PRs don't really matter that much 18:24 sfan5 haven't looked into dual wielding recently so I don't know whether the game/mod compat issue (IIRC) was solved yet 18:24 sfan5 as for settings rubenwardy has to work on it 18:24 Krock no progress on dual wield and the main menu is pretty usable already 18:24 sfan5 or if he doesn't then it doesn't get done 18:25 Krock Zughy[m]: what about the Minetest renaming? what are the next steps? poll? when would that happen? if it's an important matter for you it might make sense to draw a deadline for the submissions of your issue 18:26 Krock 5.8 is still far away and there's not too much news that would currently justify a release 18:28 MTDiscord I would recommend a private committee of devs and established community members to discuss details, purpose, and requirements for new names and to review community suggestions 18:28 Krock I suppose we already have a channel for that 18:29 Zughy[m] what greenxenith said 18:29 MTDiscord We already have a channel with both devs and established community members? 18:29 Zughy[m] this avoids both trolls and bikeshedding 18:30 Krock there's #minetest-hub, although that one's logged 18:31 nrz_ question about the main menu, the new one is always in the queue ? some concepts were very promising and more modern 18:31 ROllerozxa just throwing it out there, could #13550 make it into 5.8.0? 18:31 ShadowBot https://github.com/minetest/minetest/issues/13550 -- Debundle Minetest Game and improve UX when no game exists by rollerozxa 18:31 Zughy[m] just settings nrz_ 18:31 nrz_ oh, is there any blocker or something ? the new one with a left pane opening the other tags was sooo beautiful 18:31 Zughy[m] That'd be a great improvement UX wise (not a core dev though) ROllerozxa 18:32 MTDiscord The settings change was already merged 18:32 MTDiscord 13550 being part of 5.8.0 would be very nice 18:32 Zughy[m] nrz_: it only needs to be improved 18:32 nrz_ Zughy, thanks for the clarification 18:32 Krock there should be at least a few recommended games for newcomers 18:33 Krock it's not practical to search for the "first best" game on their own 18:33 MTDiscord ContentDB's webpage has some 18:33 MTDiscord CDB also has an API for curated recommendations which will be integrated into the menu later 18:33 ROllerozxa Krock: ContentDB puts some featured games at the top now, including Minetest Game 18:33 MTDiscord Interesting 18:33 ROllerozxa https://user-images.githubusercontent.com/60856959/248578914-54aa00c6-2e22-4941-94d3-fb9972bbfd54.png 18:34 Krock ROllerozxa: ah right. I missed that part. Seems fine to me, then. 18:34 MTDiscord Ah yes. I did see that with the new Mt client 18:34 MTDiscord A larger menu redesign still exists as a task in the pipeline, but where exactly it sits is generally unknown :] 18:35 MTDiscord Like larger cards? 18:35 nrz_ ROllerozxa: can we imagine having a link to contentDB directly in the UI ? 18:35 Krock nrz_: have a look at the PR for example 18:35 MTDiscord nrz_: Thats what the globe button is for (next to the install + button) 18:35 ROllerozxa nrz_: there are already links to packages' ContentDB pages in the content browser 18:36 nrz_ perfect Krock, 🙂 18:36 MTDiscord s20: No, an overhaul of the whole thing. But thats for a much later date. 18:36 nrz_ Rollerrozxa: for the featuring, how it's performed ? client is able to know a thing is featured ? if yes, can we imagine for example have a light green background to represent featured ? 18:37 nrz_ i see there is a text, is this processed by minetest client ? 18:37 Krock nrz_: a formspec based on json sent by the content db 18:37 nrz_ if it is, i suggest to put [featured] tag after the author name 18:37 nrz_ instead of the plain text 18:37 ROllerozxa the featuring is done server-side by putting the packages at the top 18:38 nrz_ there is no attribute in the json ? because if it's featured we should have an attribute for the client to do something about this. If there is none, possible to do that ? 🙂 18:38 ROllerozxa I'd like to make ContentDB able to give packages custom background colours so it can all be controlled server-side but that would be for another PR, I think 18:38 MTDiscord Oh, huh, it does say "Featured." as a prefix. Maybe that should be changed to [Featured] ... 18:39 nrz_ if featured is an attribute client can just highlight it properly and we can move it after author 18:39 Krock or changed to a fancy star icon in the formspec 18:39 nrz_ yeah or a star with a label 18:39 MTDiscord A star would be ambiguous and can be easily confused with favorites 18:39 nrz_ it make it easier to see for people 18:40 sfan5 Oh, huh, it does say "Featured." as a prefix. Maybe that should be changed to [Featured] ... 18:40 sfan5 yeah that looks like someone just added it to the description 18:40 sfan5 maybe a background 18:40 sfan5 or make the Features in text be colored 18:40 nrz_ about irrlicht, he hasn't been merged yet in the main source tree ? 🙂 18:41 nrz_ it seems i don't see discord NOTICE it's why i missed greenxenith, my bad, i should check my matterbridge config at a point 🙂 18:41 MTDiscord Minetest Game is already background highlighted, would it be trivial to make the featured packages also have the highlight instead of the "Featured." description text, ROllerozxa? 18:41 sfan5 irrlichtmt has not been merged 18:42 ROllerozxa greenxenith: I just removed that hardcoded check 18:42 MTDiscord Aha, could the background itself be easily re-added? 18:43 nrz_ sfan5 , any plan around this soon ? that'll permit to move irrlicht_changes in irrlicht source tree too and relink all of that properly 18:43 ROllerozxa yes, probably, if there's a "featured" field in the JSON. I was intending to save it for a later PR to be able to push this through 18:43 nrz_ luajit is mandatory now ? 18:43 Krock no 18:43 nrz_ the s_base.cpp ask me to include luajit.h 18:43 sfan5 nrz_: when the work on irrlichtmt is done 18:44 MTDiscord How trivial would it be? since it sounds like the consensus is it would be better to use a highlight if possible 18:44 sfan5 refactoring etc. 18:44 nrz_ and i had a cmake warning tolding me that i have a old luajit (i think i don't have one :D) 18:44 sfan5 clean your build environment, rerun cmake 18:44 localhost[m]1 nrz_: that's fix some irrlicht errors? Wait a sec for log... 18:44 nrz_ i think we have put an incompatible thing on luajit cmake, anyway, let's disable it 18:45 Krock nrz_: similar report https://github.com/minetest/minetest/issues/13548 18:45 sfan5 that user has a broken checkout 18:45 nrz_ Krock not the same one for me, just luajit was detected i don't know how 18:45 sfan5 the CI literally test with and without luajit, we would immediately notice if it was broken in general cases 18:45 nrz_ and cmake enabled it then failed on the include "luajit.h" 18:46 nrz_ oh okay, the luajit comes from gimp... 18:46 nrz_ luajit-2.1.0.beta3.r481.g51fb2f2c-1 > gegl > gimp 18:46 MTDiscord ROllerozxa: You could post a screenshot here before committing it to the PR in case it doesnt turn out how people expect 18:47 nrz_ nice version in stable archlinux ... 18:47 localhost[m]1 ERROR[Main]: Irrlicht: Invalid size of image for texture. 18:47 localhost[m]1 This is caused maybe by optimized textures (colors compression)? 18:47 sfan5 okay, meeting ended 18:47 sfan5 people want to talk about random issues instead it appears 18:47 MTDiscord 🙄 18:47 ROllerozxa -.- 18:48 localhost[m]1 sorry 18:48 MTDiscord You could try, instead, asking people to refrain from offtopic issues 18:48 Krock moving to past meetings 18:48 nrz_ sfan5, about the irrlicht refactoring, what do you expect on irrlicht side ? 18:48 ROllerozxa greenxenith: I'd like rubenwardy to add a featured field to the packages JSON for that, I know there's /api/welcome/ but just having it in the package list would be simpler I think 18:48 MTDiscord How are you detecting featured at the moment? 18:48 sfan5 nrz_: https://github.com/minetest/irrlicht/issues/107 18:48 nrz_ ROllerozxa: just yes ! go ahead with that 18:48 MTDiscord Or at least, where does the "Featured." prefix come from? 18:48 ROllerozxa it comes from ContentDB on the server side 18:49 MTDiscord Oh geez 18:49 ROllerozxa it gets put at the front of the description 18:49 MTDiscord Well alrighty then, I wasn't expecting that. Minetest moment right there. 18:49 sfan5 you can ask ruben to change the obviously 18:49 sfan5 the server-side should be able to check the protocol version it gets sent 18:49 ROllerozxa yes, of course 18:49 sfan5 that* 18:49 MTDiscord Later PR then for the field/highlight then, I guess 18:49 rubenwardy Please just remove the background from that PR 18:49 MTDiscord It already was 18:49 ROllerozxa rubenwardy: that's what I've done 18:50 nrz_ sfan5: about core::unicode::ustring removal, it's just time or you identify some difficult things ? 18:50 ROllerozxa no hardcoded MTG background, just debundling for now 18:50 ROllerozxa (MTG or featured...) 18:50 rubenwardy I only edit the description because Minetest releases so slowly and I want it to present on clients already 18:50 sfan5 nrz_: haven't investigated that yet 18:51 MTDiscord It was unclear that the featured label was builtin to the data, sorry for the misunderstanding there 18:51 nrz_ ok, i'll take 15 min to check what are we using for it 18:52 sfan5 we are definitely using it, it cannot be removed 18:52 sfan5 (entirely) 18:53 MTDiscord In that case it sounds like the PR is in fact ready to go, and it sounded like it already had approval before the featured appearance was discussed. But with the unceremonious death of the meeting I guess we won't be revisiting it 😛 18:54 nrz_ yes, as far i'm reading it's mandatory, and does the proper mapping 18:55 nrz_ as CPP doesn't wide any way to handle this properly 18:55 nrz_ STL i mean 19:05 sfan5 Krock: #13631 19:05 ShadowBot https://github.com/minetest/minetest/issues/13631 -- Add unit tests for isBlockInSight() by sfan5 19:09 nrz_ sfan5, i'll work a bit on cleanup the legacy ifndef define and see how we can work for the _IRR_COMPILE_WITH_OGLES1_ and etc variables. I should propose a PR this evening 19:10 nrz_ for the unittests i saw people talking about using a unittest framework, it this something we will move forward ? 19:32 nrz_ sfan5: i just noted that _IRR_COMPILE_WITH_OSX_DEVICE_ is only ifdef but not checked cmake side, no ones compiles for macox now the libirrlictmt ? 19:51 sfan5 we have CI that tests this 20:01 nrz_ sfan5: for android it's the OGLES part and we are using cmake ? if yes i should have a working thing maybe later next week (as it's less simple that expected haha) 20:48 Desour nrz_: for _IRR_COMPILE_WITH_OSX_DEVICE_, see https://github.com/minetest/irrlicht/blob/83699a835df2c7363ae856ee128fb2409743f65d/source/Irrlicht/CMakeLists.txt#L118 21:06 nrz_ Desour, I'm working on this, it's why i ask 21:13 Desour it looked like you didn't see that line, that's why I showed it you :)