Time Nick Message 00:04 MillersMan hmmmm, paramat: I'm currently hacking again on #4061 00:05 ShadowBot https://github.com/minetest/minetest/issues/4061 -- Add loaded blocks to transforming_liquid-queue similar to updateLiquid by MillersMan 00:06 MillersMan My fix will access the block above and below in addition to the loaded block and might also need to access to 12 further under some circumstances 00:07 MillersMan Currently I'm fetching those 12 additional blocks for each loaded block, but I think that this might be a costly operation so I'm thinking about lazy fetching. 00:08 MillersMan Only problem: I have no idea how I might differentiate between a not yet fetched block and a not yet loaded block that was returned as NULL when fetching, any ideas? 00:09 MillersMan My ideas so far were to mark not loaded blocks with ~0, which is obviously an ugly hack which might not be ported to all architectures 00:11 MillersMan Or to create a special block which resides outside the valid game-range and consists of only CONTENT_IGNORE nodes (which removes the need for an extra check whether the block exists) 00:56 paramat MillersMan what do you think about my suggestion in the last comment? that would fix the main problem while being simpler 01:12 Wuzzy Jeija asks: “shouldn't the cave generator only add caves when emerging new mapblocks?” 01:12 Wuzzy https://github.com/Jeija/minetest-mod-mesecons/issues/295#issuecomment-262047199 01:13 Wuzzy Jeija wonders if it is strictly neccessary to set is_ground_content=false for non-stone/dirt/etc. nodes 01:14 Wuzzy any idea how i should answer? 01:14 garywhite ~seen garywhite 01:14 ShadowBot garywhite: I saw garywhite in #minetest-dev 1 second ago saying "~seen garywhite" 01:14 garywhite ~seen celeron55 01:14 ShadowBot garywhite: I saw celeron55 in #minetest-dev 2 days, 13 hours, 30 minutes, and 19 seconds ago saying "it's like a suggestion, but definitely bad" 01:17 paramat large caves generate out past the edge of each mapchunk by up to 16 nodes, into pre-generated mapchnks 01:18 paramat any mesecons added in 'on generated' are at risk 01:19 paramat but i expect that doesn't happen much 01:19 paramat (mapchunks) 01:20 paramat still worth making them not ground content 01:28 paramat essential actually, schematic structures containing mesecons may be added in 'on generated' 01:42 cheapie So far, my attempt at sort of working around the dropped-players issue (commenting-out lines 2184-2190 of environment.cpp, more or less) hasn't caused any noticeable issues, and that's after rather extensively playing a world that used to have the dropped-players issue *very* frequently. 01:42 cheapie I'd still like to see some other people testing that, in case I'm missing something... 01:43 cheapie The bug in question is #4067, and the world in question can be downloaded here: http://daconcepts.com/vanessa/hobbies/minetest/worlds/Creative_World.tar.bz2 01:43 ShadowBot https://github.com/minetest/minetest/issues/4067 -- Canceling: No player for peer_id when too much objects on the map 01:53 Wuzzy thanks for the info :) 02:18 cheapie PR created for the above, including a more detailed explanation: https://github.com/minetest/minetest/pull/4802 02:52 paramat game#1404 will merge soon before conflicts arise. please don't merge any texture changes before this 02:52 ShadowBot https://github.com/minetest/minetest_game/issues/1404 -- Convert many indexed-with-alpha textures to RGBA by paramat 02:59 Wuzzy hmmm 03:00 Wuzzy I feel the new boost cart in Minetest Game is implemented poorly, to be honest 03:00 Wuzzy sorry :( 03:00 Wuzzy it is so ... i dont know the word 03:00 Wuzzy juddery? 03:01 Wuzzy i mean the cart does not run fluidly 03:01 Wuzzy instead it seems to jump around while driving. very unpleasant to watch 03:02 Wuzzy this is very odd since when I replace the built-in carts mod with the original boost_cart, the boost cart works just fine 03:02 Wuzzy maybe bad merge? 03:03 Wuzzy also, why on earth did you replace the sound? it is so quiet, it can barely be heard. the new texture looks fine, however 03:06 paramat please do give feedback, maybe open an issue, carts will need refining before release 03:06 paramat and yeah the sounds could be made a little louder 03:11 paramat perhaps lua lag is having a larger effect on merged carts for some reason 03:22 paramat they ran smooth for me but i had no other mods running 03:24 sofar I wanted the sound to be louder, but paramat is making all sounds softer :) 03:24 sofar Wuzzy: any chance you could video that? 03:24 sofar the jittery thing you're talking about? 03:25 sofar the movement code is pretty much exactly like boost_cart 03:25 Wuzzy weird 03:25 Wuzzy no, sorry 03:25 Wuzzy paramat: why not just using the original sounds? bad license? 03:26 Wuzzy but *maybe* i could live with the new sound if it were louder :) 03:26 sofar original sounds? there were no sounds in boost_cart 03:26 sofar https://github.com/SmallJoker/boost_cart 03:26 sofar no sounds folder there 03:30 paramat ugh now i'm getting visual bugs with mip-mapping even with textures converted to RGBA, perhpas my GPU just can't handle mipmap 03:30 paramat (perhaps) 03:30 Wuzzy oops 03:30 sofar lol 03:31 Wuzzy sofar: Well maybe I was thinking of the sound used in @test@ 03:31 kaeza revert the revert! 03:31 paramat so now i have no idea what to do about game#1397 03:31 ShadowBot https://github.com/minetest/minetest_game/issues/1397 -- Visual regression after indexed images merge 03:31 kaeza paramat, what GPU? 03:32 paramat intel integrated 'haswell desktop' 03:33 garywhite Will someone explain please what game#1397 does exactly? 03:33 ShadowBot https://github.com/minetest/minetest_game/issues/1397 -- Visual regression after indexed images merge 03:37 paramat beginning to suspect that bug is unfixable for certain GPUs 03:44 paramat so, i'll fix game#1402 instead 03:44 ShadowBot https://github.com/minetest/minetest_game/issues/1402 -- debug spam with "nodeupdate: deprecated" message 03:46 paramat and test the new ceiling torch 03:48 garywhite paramat: What happens after the feature freeze on 7 December? 03:48 paramat testing for bugs 03:49 paramat and fixing any remaining issues 03:49 garywhite So basically you just play around with worlds in the game and report on any issues? 03:55 paramat the freeze avoids any bugs creeping in at the last moment, bugs can sometimes take a while to be discovered 03:55 paramat that's why new features are frozen 03:55 paramat torches are lovely +1 04:10 cheapie Speaking of bugs, any thoughts on #4802 yet? 04:10 ShadowBot https://github.com/minetest/minetest/issues/4802 -- Don't clear excess entities on mapblock save (possible fix for #4067) by cheapie 04:48 paramat game#1405 04:48 ShadowBot https://github.com/minetest/minetest_game/issues/1405 -- Update 'nodeupdate()' to 'check_for_falling()' API by paramat 04:48 paramat will test and hopefully merge later 05:01 sofar glad you like the ceiling one... I have my doubts on the aestetics myself, but this is the best I can think it should be 05:07 paramat it's fine 05:10 sofar tested the whole conversion again, made the lbm name a bit shorter as well 05:10 sofar should be ready for merge 05:10 paramat good 05:10 sofar wait 24h and then merge? 05:12 paramat yes maybe a good idea to be careful due to the critical LBM conversion stuff 05:15 sofar well, I kept all the nodes as WALLMOUNTED 05:15 sofar that makes the conversion really simple 05:15 sofar looks fine in most texture packs 05:19 paramat i can't really judge if we should wait, you can decide when it's safe to merge 05:22 sofar give it 24h, then we can merge. Always nice to give people a time to react. 05:24 sofar merged game#1403 :) 05:24 ShadowBot https://github.com/minetest/minetest_game/issues/1403 -- Make smoke puff sound mono by sofar 05:27 sofar game#1314 has been sitting too long 05:27 ShadowBot https://github.com/minetest/minetest_game/issues/1314 -- Use travis' container-based infrastructure by sfan5 05:37 Hijiri what if the deep learning meme was applied to anticheat? 05:38 Hijiri I'm sure it's been done in other games 05:44 sofar the problem is that you can't define "cheating" easily 05:44 sofar is moving too fast cheating? -> problems with lag 05:44 sofar digging too far? cheaters will dig less far 05:44 sofar xray? cheaters will mine randomly 05:47 Hijiri it might be possible for movement cheats if you could get a representative learning set (different physics overrides, lag levels, etc.) 05:48 Hijiri though that part might be hard 05:48 Hijiri getting sets of data for cheated movement, since most players' movements are not going to be cheating 05:48 sofar you know 05:48 sofar that's the stupidest use of deep learning we can do in minetest 05:48 sofar I've got a better idea 05:48 Hijiri It seems like a pretty good place to use it 05:49 Hijiri anticheat we haven't found good heuristics yet for 05:49 Hijiri but surely one exists, because you can look at a cheater and tell if they are cheating (sometimes) 05:49 Hijiri for flying, at least 05:49 sofar how about we use deep learning to drive npc's to do things like farm, hunt players, build structures? 05:50 Hijiri That seems like a good use too, but I think anticheat is actually a very good place for machine learning 05:50 Hijiri precisely because it's hard to define on paper what cheating is 05:51 Hijiri not to say that it's a more interesting use, but it was the topic I brought up originally 05:53 paramat will merge game 1405 in a moment 05:54 Hijiri there might be good non-learned heuristics for flying already, that just aren't published because they are implemented as part of proprietary games 05:56 paramat merging 05:58 paramat complete 09:35 Zeno` I scared the kwoleker away 09:37 Zeno` +1 for #4800 09:37 ShadowBot https://github.com/minetest/minetest/issues/4800 -- Fix superflous shader setting updates by ShadowNinja 09:38 Zeno` valgrind is clean 09:38 Zeno` I'm not going to performance profile it because I don't think it's something that valgrind will pick up anyway 09:38 Zeno` valgrind/callgrind 09:39 Zeno` there are still leaks, *BUT* I think (99.9% sure) that those are irrlicht's fault 09:39 Zeno` I'm fine with it 09:39 Zeno` and they're not leaks that will keep increasing anyway 09:40 Zeno` just unfreed memory 09:40 Zeno` (allegedly) 09:40 Zeno` massif shows nothing unusual 13:46 Zeno` will merge #4800 shortly 13:46 ShadowBot https://github.com/minetest/minetest/issues/4800 -- Fix superflous shader setting updates by ShadowNinja 13:59 Fixer Zeno`: needs second approval i think 13:59 Zeno` ShadowNinja is a core dev, I am, that's two approvals 14:00 Zeno` that's how celeron55 explained to me it works 14:01 Zeno` if it didn't work that way things would be even slower than they are now :) 14:04 nrzkt Zeno`: approved too, then three approvals 14:05 Zeno` even if that wasn't the case, the PR has been debated to death 14:05 Zeno` it's time to merge. Doing so now :) 14:08 Zeno` I don't agree with everything in the PR, but I cannot think of a better way 14:09 Zeno` so I could no longer in good faith not agree to merging it 14:17 Zeno` not disagree lol 14:18 Fixer time to test 14:18 Zeno` "not object" is clearer 14:19 Zeno` err "object" 14:19 Zeno` darn English is hard! 14:39 Zeno` #4803 14:39 ShadowBot https://github.com/minetest/minetest/issues/4803 -- Update tab_credits.lua by Zeno- 14:40 Zeno` he has done *nothing* for the current release except cause malcontent between core devs and generally disrupt development 14:40 Zeno` those actions are not something that deserve credit 14:41 Zeno` before he came along we worked as a TEAM 14:42 Zeno` he cannot seem to grasp how to work with a team and there I don't think he belongs being there 14:42 Zeno` s/there/therefore 14:43 Zeno` besides that, he said he wants nothing to do with us anyway 15:24 Zeno` made a merge 15:24 Zeno` please update repos 15:59 Zeno` how can we review all these PRs? 16:00 Zeno` is #4837 ok to merge? 16:00 ShadowBot Zeno`: Error: Delimiter not found in "HTTP Error 404: Not Found" 16:00 Zeno` err #4737 16:00 ShadowBot https://github.com/minetest/minetest/issues/4737 -- Add minetest.close_formspec() to close formspecs from the server side by orwell96 16:01 Zeno` added my +1 16:01 Zeno` will merge soon 16:02 Zeno` the quantity of PRs is distressing me :/ 16:03 Zeno` things are out of control 16:03 Zeno` there is no way a team as small as ours can catch up with this 16:05 Zeno` can someone ask if lhofhansl wants to be a dev? 16:05 Zeno` his PRs are good 16:05 Zeno` https://www.linkedin.com/in/larshofhansl 16:05 Zeno` we really need more devs :( 16:06 Zeno` just to handle @&$*&@* code reviews if anything 16:07 Zeno` if this was a commercial project I'd hire him 16:07 Zeno` if he wants in, please add him 16:08 Zeno` up to him of course. Maybe he doesn't want to be 16:08 Zeno` celeron55 ^^ 16:09 nrzkt Zeno`: warning he works at salesforces, he will buy MT and sell it for customers :D 16:09 nrzkt xD 16:09 Zeno` lol, nah :P~ 16:09 Zeno` and who cares if he does anyway (so long as license is complied with)? 16:10 nrzkt :p 16:11 Zeno` *IF* he wants to sell minetest, he can already 16:11 Zeno` we're better off having him as part of the team no matter 16:11 Zeno` to help with this backlog of stuff 16:12 Zeno` for a start 16:12 Zeno` every PR of his has been pretty much faultless 16:14 Zeno` hmmmm ^^ 16:16 millersman hi, how stable is mingw crosscompilation? 16:17 Zeno` the official Windows minetest versions are cross-compiled using mingw 16:17 millersman Zeno`: Is this true for 64bit also, got some errors when I tried last time 16:17 Zeno` not sure for 64-bit 16:18 Fixer millersman: i do it and use all the time, works fine 16:18 Zeno` sfan5 would know 16:18 Fixer millersman: crosscompiling on debian via mingw64 to windows 64bit 16:19 Zeno` millersman, in general sfan5 makes (afaik) the Windows builds (well, his scripts do) 16:19 Fixer millersman: it is as stable as msvc one imo 16:19 millersman ok, i'll pull and try again, i'm compiling in an 64bit arch vbox for windows 64 16:20 Zeno` millersman, if you encounter an error can you post details as a bug report on github? 16:20 millersman ok, i'll do 16:21 Zeno` that'd be great :) 16:22 garywhite Is there a build for Debian 32? Cause I'm setting up a VM in Windows and it's based on Debian 32-bit 16:22 Fixer Zeno`: i did some testing of 4800 on 9 scenes: on average boost is 38% (ranging from 24% to 85% depending on scene), it provides bigger boost in heavy scenes to me, I'm glad it got merged 16:22 millersman in other news: got reflow almost working when my system suddenly froze, but i should be able to recode it from memory 16:22 Zeno` Fixer, very nice improvement! 16:22 nrzkt Fixer: it's normal memory action vs vector + select + ugly string comparision = ouch 16:23 Fixer median is 41% 16:23 Fixer as was reported by other people 16:23 Fixer very nice help for heavy scenes 16:24 Fixer so on my PC, with default settings (which is just basic shader on), avergae penalty in those scenes is 12%, before patch - 62% 16:25 Fixer shaders are much more usable now 16:27 Fixer in short, just enable fps and you will lose just 10% of fps, not 30-50% like before patch 16:27 Fixer enable shaders* 16:30 Fixer in short: it is awesome 16:33 Zeno` I'll revert it then? 16:34 Zeno` will revert #4800 in ten minutes. It's too powerful 16:34 ShadowBot https://github.com/minetest/minetest/issues/4800 -- Fix superflous shader setting updates by ShadowNinja 16:35 Zeno` lol 16:40 Fixer paramat said there another thing remains that can potentially increase shaders performance some more, it is disabling some feature RBA enabled earlier for all shaders (vertex spaces or how it is called) in cases when bumpmapping and parallax is not used I think , don't remember details 16:42 Fixer can provide up to 20% fps boost iirc 16:42 jin_xi tangent space 16:42 Fixer yes 16:42 Fixer tangent space 16:43 Fixer it lowers performance, but is used only for some shaders 16:43 Fixer it is used for everything now I believe 16:53 millersman Zeno`: When linking I get undefined references to some symbols starting with leveldb:: any idea? 16:54 Krock oh cool. Someone else with the same problem as me 16:54 Krock with MSVC, 32-bit 16:55 millersman here with MinGW 64bit cross compilation in arch 64bit 16:56 millersman Krock: is there any older version that worked for you? 16:56 Fixer millersman: does not compile? 16:56 Fixer millersman: maybe disable leveldb via cmake 16:56 Krock millersman, I gave up trying two years ago :3 16:56 Krock It never worked 16:57 millersman Fixer: Thanks, will try it 16:57 Fixer i have it enabled though 16:57 Krock disabled it and included postgres later instead 16:57 millersman leveldb is the default backend, right? 16:58 Fixer no idea 16:58 Fixer everything compiles for me 16:58 Wayward_One sqlite 16:58 Krock millersman, no, sqlite3 is 16:58 Fixer i use libs provided by sfan 16:59 millersman ok, so disabling leveldb won't interfer with opening games created with the default version, good 17:02 millersman Fixer: Do you still have a copy of the reflow test-world? My copy is on the broken system 17:02 Fixer did I shared it with you? 17:03 millersman yes, you did when i started work on the reflow fix sometime in march or april, was a zip containing fragments of a bigger world 17:03 Fixer hmmm 17:03 Fixer i have it 17:04 millersman i used it as the the test for gabs and missing flow updates 17:04 millersman *gaps 17:05 Fixer yes, i have it now 17:06 Fixer millersman: do you want it? 17:07 millersman yes, would be nice to have a good test-case 17:09 Fixer where i can upload it? it is 40mb 17:11 Fixer maybe i can split it up into many zips and post on github 17:12 millersman i don't remember how you shared it last time, i don't have anything except dropbox 17:15 red-002 https://transfer.sh ? 17:17 Fixer red-002: ty, will try now 17:20 Fixer millersman: try this https://transfer.sh/3o6pn/waterflow.zip 17:23 Fixer millersman: is it working for you? 17:27 millersman Fixer: Download worked, have to setup minetest to test it 17:27 Fixer if archive is correct, it is cool, added to bookmarks 17:27 millersman And compilation worked also 17:31 sfan5 millersman: seems like you found the correct build scripts 17:31 sfan5 the problem you encountered is already described in an issue here https://github.com/minetest/minetest/issues/3026 17:35 Fixer ha, debians one is probably older 17:38 millersman sfan5: unfortunately it doesn't start, looks like functions for the c++standard lib aren't found, e.g. thread::start_thread or so 17:38 Fixer millersman: did you copy libgcc_s_seh-1 and libstdc++-6 to your /bin ? 17:38 sfan5 ^ 17:39 Fixer those files can be found in mingw64 libs i think 17:39 sfan5 and i'm pretty sure it avoid c++11 threads on windows on purpose 17:40 Fixer /usr/lib/gcc/x86_64-w64-mingw32/4.6/ or so 17:40 sfan5 nvm it actually tries using c++11 threads 17:41 Fixer libgcc_s_seh-1.dll or LIBGCC_S_SJLJ-1.DLL / and this one libstdc++-6.dll 17:41 Fixer http://arrayfire.com/cross-compile-to-windows-from-linux/ 17:47 milllersman yeah, it runs... took me just some time to locate the DLLs on my system where they are in /usr/x86_64-w64-mingw32/bin/ 18:45 milllersman Fixer: The world you uploaded looks different from what I had for testing, the other one was more like a fragment from a very active server. But the lavafalls should work too 18:50 milllersman Fixer: And also found the log of our original conversation with the link to the world I was looking for 18:54 Fixer nice 20:27 paramat i'd also like to have lhofhansl as a core dev if he wants to 20:30 red-001 would anyone have time to review #4472 ? 20:30 ShadowBot https://github.com/minetest/minetest/issues/4472 -- Make serverlist searchable. by red-001 20:38 garywhite Does anyone know where I can get the Minetest packages in .deb format? 20:39 red-001 https://launchpad.net/~minetestdevs/+archive/ubuntu/stable ? 20:40 garywhite ok, lemme log in via Linux now so I can follow the link 20:41 garywhite Got the link 20:42 paramat nore sfan5 (ShadowNinja rubenwardy) 3D torches is ready for merge game#1325 the new ceiling torch has been added please review, can we merge now? 20:42 ShadowBot https://github.com/minetest/minetest_game/issues/1325 -- Merge 'torches' mod as default:torch. by sofar 20:58 garywhite Unfortunately, the repository will not install 21:10 milllersman paramat: there's no need to do a partial implementation for #4061, got it almost working... before my system broke... so i'm currently redoing it on the laptop 21:10 ShadowBot https://github.com/minetest/minetest/issues/4061 -- Add loaded blocks to transforming_liquid-queue similar to updateLiquid by MillersMan 21:10 paramat ok 21:11 milllersman i hope to have it ready for commit after this weekend 21:19 Fixer fingers crossed 22:30 paramat hi milllersman could you fix the join / leave spam coming from your connection? 23:20 Fixer paramat: just disable those pesky join/leave messages 23:21 cheapie Could anybody please take a look at this line and see where I should split it? I've been told it's too long, but I don't know enough of C++ or MT's style guidelines to determine where to split it: https://github.com/minetest/minetest/pull/4802/files#diff-baa4d2d0ffafdec2b86baf065fba252bR2190 23:48 rubenwardy ga,e 23:48 rubenwardy game#1179 rebased 23:48 ShadowBot https://github.com/minetest/minetest_game/issues/1179 -- Add simple fast inventory by rubenwardy 23:49 rubenwardy sofar ^ 23:49 paramat cheapie split at 80 columns (if tab size is set to 4) or a max of 90 columns if 90 makes it significantly more readable 23:49 * cheapie sets his tab size to 4 and looks again 23:50 cheapie OK, I see your point. Where should I split it? I don't know enough C++ to tell where I can and can't. 23:52 paramat after '!=' 23:52 cheapie OK, and am I supposed to have the next line indented in that case? 23:52 paramat then 2 indents for the next line 23:52 cheapie OK, thanks. 23:53 paramat to distinguish from the lines following 23:53 cheapie I'll push it in a second. 23:53 paramat and a space after the 'if' 23:55 cheapie OK, here's what I have now: https://github.com/cheapie/minetest/blob/cd6eae51b76bc08e0a6f682d3cf0ae3e592e84ed/src/environment.cpp#L2190 23:55 paramat fine, but space before the final { on L2191 23:57 cheapie Done: https://github.com/cheapie/minetest/blob/4e28e9d5347a321c477e3473b62e8a4c4b5ebc84/src/environment.cpp#L2191 23:57 paramat and see http://dev.minetest.net/Code_style_guidelines 23:57 paramat yeah good 23:58 cheapie That looks handy... 23:58 * cheapie bookmarks that