Time Nick Message 00:00 Bratmon I played on a minecraft server plugin called Naval Battlezone with a similar idea. 00:00 Bratmon The implementation was god-awful, but it still was really fun. 00:04 Bratmon The minimum viable product on them isn't actually that much. 00:05 Bratmon Just nodes that know that they're a boat, and then all the nodes move together. 00:21 hmmmm well first you need to make each voxel area entity its own irrlicht scene node 00:21 hmmmm then you have to define a way to store these in the map database 00:22 hmmmm then you have to define separate mesh making code for it 00:22 hmmmm then you have to modify the protocol to work with these 00:23 hmmmm were you volunteering for that, bratmon? 00:25 Bratmon If you're a lazy bastard, you can get half way there by just making pseudo-VAEs 00:25 Bratmon That are just nodes in the world with some metadata. 00:26 kaeza someone already did (kinda) 00:26 Bratmon Where? I was unaware of this. 00:27 jin_xi stu did that i believe but i never tried 00:29 kaeza it works pretty well, but the disadvantage is lack of non-aa collision boxes 00:29 Bratmon aa? 00:29 jin_xi axis aligned 00:30 Bratmon That makes sense. 00:34 kaeza https://forum.minetest.net/viewtopic.php?f=11&t=8059 00:36 Bratmon Thanks. kaeza 00:36 Bratmon That's exactly what I was thinking of. 02:37 paramat 'Therefore a single noise parameter table can be used to get 2D or 3D noise, when getting 2D noise spread.z is ignored.' < correct hmmmm ? 02:38 hmmmm that is correct 02:38 paramat i've seen today's biome drama and corrected my misunderstanding about spread.z not needing to be 1 02:38 hmmmm biome drama? 02:40 hmmmm lol before some guy started spazzing out before... "OMG MAPGEN V7 IS SOOOOO MESSED UP! IT WAS FINE UP UNTIL A COUPLE DAYS AGO AND IT NOW EVERYTHING IS JUST TOTALLY MESSED UP" 02:41 hmmmm :/ i forgot to remove the manual noise transforms to biome noise when i added auto-transform to the perlinMap* functions 02:42 paramat yep saw that. beginning to realise you don't test much of your work heh 02:43 hmmmm not during that flurry of activity 02:43 paramat that's understandable 02:43 hmmmm speaking of biomes 02:44 hmmmm i'm thinking of adding the biome infrastructure to v6 02:44 paramat gosh 02:44 hmmmm :D 02:44 hmmmm "NOOO MAPGEN V6 WAS FINE UP UNTIL A COUPLE DAYS AGO" 02:44 hmmmm "NOW IT'S ALL ROCK" 02:44 hmmmm "BRING BACK THE OLD VERSION!" 02:45 paramat lol 02:45 paramat place decos is in there, i was thinking of removing that 02:45 hmmmm i think i'm going to define some default biomes in builtin 02:45 hmmmm getting real tired of this BS 02:46 paramat well i'm working on default biomes already https://github.com/paramat/minetest_game/tree/mgv5biomes 02:46 hmmmm cool but that's part of minetest game 02:46 hmmmm so not really default 02:46 paramat oh i see 02:47 paramat just something really basic and not stone then 02:47 hmmmm plains 02:47 hmmmm i should do this before the release 02:48 paramat i saw a video review of minetest, only stone v7 was used 02:48 hmmmm :| 02:48 hmmmm "what a crappy game" 02:48 paramat "seemed bleak" < actual comment on IRC 02:48 hmmmm hah 02:49 paramat "some kind of hardcore survival?" 02:51 paramat i'm also working on improving the biome api, i removed 'dust water' because everyone uses that for ice, but that creates holes in ice under overhangs/floatlands 02:51 paramat replaced with 'water top' 02:51 hmmmm really? 02:51 paramat could also add 'water top depth' for ice depth 02:51 hmmmm i'm planning on nuking the current biome system because it seriously blows 02:52 paramat well see my improvents first https://github.com/paramat/minetest/tree/biomeapinodes 02:53 paramat i have ideas and would like a chance to improve it 02:54 paramat solved the 'surface node underwater' problem when under overhangs 02:55 paramat but also kept height min/max so vertically layered biomes still work, would work for stacked realms 02:55 paramat so yes i checked everyone who is using the biome api and 'dust water' was only used for icesheet 02:56 paramat my improvements combined with my MTgame biomes create an impressive result 02:58 paramat what's good about the current biome api is it uses 2D noise for heat/humidity but can also cope with vertically stacked biome systems. i consider it a fairly good system, just needs some work 03:02 hmmmm my idea behind the whole water and 'water dust' was maybe it could be a fog node for swamp biomes 03:02 hmmmm or maybe radioactive dust around a radioactive liquid biome 03:03 hmmmm etc. 03:07 paramat aha, i was wondering what it was for, could be re-added 03:09 paramat so perhaps instead of yet another trashing we can improve what we have, organically evolve it in little steps 03:10 paramat i remember a while back in IRC you gave me the job of sorting out biomes, i've put a lot of work into it 03:11 Tesseract v5 and v7 should be marked as (experimental) and (incomplete) before release. 03:11 paramat yes 03:11 paramat i need until 0.4.12 to make tree schems, complete the biome system and improve the biome api 03:14 paramat hmmmm, if you let me know what you don't like about the biome api i/we can improve what we have. trashing it means a long wait after which you might get fed up with your new system and trash it again 03:20 paramat as i wrote in IRC recently we could have multiple biome apis, insisting on one system just causes unnecessary grief for ourselves 03:27 paramat any biome api inherently results in limitations in what it can do, therefore insisting on one biome api for all mapgens then rules out a vast amount of other possibilities 03:29 paramat those lost possibilities then make us feel bad and make us consider trashing what we have for a better single biome api 03:30 paramat pointless grief =) 03:32 paramat it's a bit like insisting on only one mapgen, it would be hell trying to choose one, and working on one only leads to numbness through over-familiarity 03:34 paramat we could leave v6 as it is. v5 and an improved v7 could be current biome api. v8 would use your new biome api 03:34 paramat or perhaps even make each biome api interchangeable across multiple mapgen base terrains 03:48 MinerDad Is adding parts to the formspec kosher? I'm looking at https://github.com/minetest/minetest/issues/1374 and I think giving mods an ability to set a prediction for put/take would make it possible 03:50 MinerDad Either change formspec or add a new parameter to the node to indicate this per list 03:53 hmmmm paramat, what i had in mind was to move generateBiomes() to BiomeManager and have it selectable 03:56 paramat so it could be disabled, sounds good 03:57 hmmmm at first i planned to have an mgv6 spflag for "use system biomes" 03:57 hmmmm might make a new mapgen param for "biome manager type" 03:57 hmmmm and each mapgen has its default 03:58 hmmmm v6's default will be "none", v5 will be "paramat's", and v7 will be "hmmmm's" 03:58 hmmmm and you can switch between all of them 03:58 hmmmm yes? no? 04:00 paramat yes good stuff 04:07 paramat i really look forward to seeing v8 and your new biome api ideas, i feel people will be much happier knowing this wouldn't involve losing what we have 04:32 hmmmm http://i.imgur.com/bsU7iam.png 04:32 hmmmm blob ore ^ 04:32 hmmmm this is supposed to replace dirt, gravel, and clay 04:32 * paramat looks 04:33 paramat what i'm doing with the current biome api is a backwards-compatible improvement, the practicality of this is keeping the many players currently using it happy, only a few changes to their biome definitions will be needed 04:33 hmmmm well 04:33 hmmmm it doesn't need to be stable at all 04:33 hmmmm biomes are not officially supported 04:36 paramat yeah 04:38 gregorycu I don't know 04:38 gregorycu The floating coal just doesn't seem realistic to me 04:39 hmmmm lol 04:41 Bratmon gregorycu, Are you kidding? I'm looking out my window right now, and I see nothing but giant mountains of floating coal. 04:41 gregorycu Wow 04:41 gregorycu The question is, do you need to see a doctor, or do I need to see a geologist? 04:42 Bratmon I think both. 04:42 hmmmm how bout a proctologist 04:42 gregorycu That sounds dirty 04:42 gregorycu It was more dirty than I expected 04:43 Bratmon I don't like where this is going. 04:43 hmmmm me neither. it sounds like off-topic banter that belongs in #minetest 04:43 Bratmon That's fair. 04:44 gregorycu You started it! Anyway.... 04:46 Zeno` pumice floats 05:16 gregorycu I'm trying to learn how to use an ABM 05:16 gregorycu function(pos, node, active_object_count, active_object_count_wider), pos is the position, node is the node object 05:17 gregorycu How do I get a handle to the object the represents the node type 05:17 gregorycu (The thing that is registered with register_node) 05:23 gregorycu minetest.registered_nodes is the thing you use 05:51 Zeno` #1986 05:51 ShadowBot https://github.com/minetest/minetest/issues/1986 -- Fix visual_scale for NDT_PLANTLIKE being set too small by Zeno- 08:55 Zeno` Any objections to #1986 being merged? 08:55 ShadowBot https://github.com/minetest/minetest/issues/1986 -- Fix visual_scale for NDT_PLANTLIKE being set too small by Zeno- 09:05 Megaf Zeno`: I've noticed no regressions in that 09:05 Zeno` Nor I. It's a simple fix really 09:22 Zeno` I'm going to push a simple fix to stop annoying warning. Changing https://github.com/minetest/minetest/commit/67063ed59f91d2feae5011ca0c3767dc182bf38d#diff-5a3527c6dddd0430cefor (s32 i = 0; i != sx * sy; i++)6d04404ea096baR75 to 09:23 Zeno` great paste zeno! 09:23 Zeno` changing https://github.com/minetest/minetest/commit/67063ed59f91d2feae5011ca0c3767dc182bf38d#diff-5a3527c6dddd0430ce6d04404ea096baR75 to for (s32 i = 0; i != sx * sy; i++) 09:55 Zeno` will merge 1986 in about 30 minutes 10:15 casimir #1851 was closed, but random-input is still broken. 10:15 ShadowBot https://github.com/minetest/minetest/issues/1851 -- Cleanup updateCameraDirection and fix random input not working by Zeno- 10:17 Zeno` in what way is it broken? 10:21 casimir It always looks into the same direction. 10:21 casimir I tested your pr and it fixed it. 10:21 Zeno` which PR? 10:22 casimir #1851 10:22 ShadowBot https://github.com/minetest/minetest/issues/1851 -- Cleanup updateCameraDirection and fix random input not working by Zeno- 10:23 Zeno` it's already merged though 10:25 casimir That is strange. 10:27 Zeno` indeed :) 10:42 casimir From what I see it was not merged, but I will take a closer look at this issue tomorrow when I have the time. 10:45 gregorycu Hello all 10:45 gregorycu What is the target C++ version? 10:48 Zeno` 98 10:49 gregorycu Ok 10:49 gregorycu Anyway, you got my PMs? 10:49 gregorycu 38% -> 25% CPU 10:49 gregorycu That's in debug, I'll have to try release 13:45 Zeno` hmmmm, no 13:45 Zeno` just no 13:45 Zeno` I did not change that to a signed type for the reason you're suggesting at all 13:46 Zeno` I changed it because the literal is an int and also that int is the fastest type for a loop (in this case) 13:48 hmmmm unsigned int is slower than int?? 13:48 Zeno` signedness did not even enter my mind 13:48 hmmmm that comment was based on both of your previous two commits 13:48 Zeno` u16 is not necessarily the native type of the target processor 13:49 hmmmm yes, but unsigned int is not necessarily u16 13:49 Zeno` yes 13:49 Zeno` so int, in this case, is the correct type 13:49 Zeno` As for the previous commit... well, I hate warnings 13:50 hmmmm as for the compiler warning you silenced; i didn't even see that because my version of gcc emits a crapton of spurious warnings about conditions always evaluating to true/false due to type range limitations.. so it's easy to miss 13:50 Zeno` but this is not related 13:50 hmmmm i was holding out on that when somebody brought it up because it requires fixing the signedness of the coordinates which is an issue with a lot of my older code 13:50 Zeno` hmmmm, the previous commit was not suggesting you did anything wrong 13:50 hmmmm yeah but it's not a fix 13:50 hmmmm it just silences the warning 13:51 Zeno` not in a way that can fail though 13:51 hmmmm correctness is important 13:51 Zeno` it is correct 13:51 hmmmm using int for an array index variable is not correct 13:51 Zeno` x is signed, y is signed 13:51 hmmmm arrays should be accessed using unsigned types always 13:51 Zeno` hmm, of course it is when the condition being checked is... an int! 13:51 Zeno` the literal 4 is of type int 13:52 hmmmm yes... so? 13:52 Zeno` so int is correct 13:52 hmmmm no, it's not correct because you're using the variable 'i' as the index to an array 13:52 hmmmm the constant 4 gets promoted to 4u anyway 13:52 Zeno` why is that not correct? 13:53 hmmmm because it's possible for the array index to be negative 13:53 Zeno` array[-1] is most certainly valid 13:53 hmmmm it certainly compiles 13:53 hmmmm but that doesn't mean it's good practice 13:53 Zeno` array indices CAN be negative 13:53 Zeno` array[-1] is *(array -1) 13:53 hmmmm they can be 13:53 Zeno` of course 13:53 hmmmm but that doesn't mean it's good 13:54 hmmmm i can think of a million things that work just fine but are bad practice 13:54 Zeno` complain to the C and C++ standard's committees not me then :P 13:54 Zeno` I can think of a million reasons where negative indices are good 13:54 hmmmm in fact certain compilers will give warnings for signed indexes 13:55 gregorycu What is the type of array? 13:55 Zeno` they should not because it's a valid thing. If a compiler warns about that then it is non-compliant 13:59 hmmmm it's a warning 14:00 Zeno` yes 14:00 hmmmm a warning says something could possibly be dangerous 14:00 Zeno` and too many warnings lead to people ignoring warnings 14:00 hmmmm not that it's non-conforming to the standard 14:00 hmmmm using strcpy() produces warnings under a lot of compilers; that's part of the standard though 14:01 hmmmm and what about while (foo = bar()) { 14:01 hmmmm that produces warnings 14:01 hmmmm it's still 100% correct C 14:01 Zeno` of course it does 14:01 Zeno` because it is 9 times out of 10 an error 14:01 hmmmm and using signed integral types for array indexes? 14:02 hmmmm not necessarily an error but could be improved 14:02 Zeno` I'm serious... *(array - n) is common 14:03 Zeno` but this is not what we're talking about here 14:03 hmmmm array notation implies different things to the person looking at the code 14:03 hmmmm vs. pointer arithmetic notation 14:03 Zeno` they are the same 14:04 Zeno` but anyway, the real point is that u16 is almost certainly not the native word size of any CPU since about 20 years ago 14:04 Zeno` so there is an implicit cast to s32 (or whatever it may be). Of course -O3 will fix that, but *shrug* 14:05 Zeno` if the rhs of a condition is int then the lhs should be as well 14:05 Bratmon I think you two might be bikeshedding at this point. 14:05 hmmmm as far as i'm aware, the only speed guarantees are made with int_fastN_t 14:06 Zeno` hmmmm, correct 14:06 hmmmm where does it say that int is the fastest type on a given platform 14:06 hmmmm so your point is kinda moot 14:06 Zeno` not really 14:07 Zeno` I really don't believe we're arguing about this 14:07 Zeno` change https://github.com/minetest/minetest/commit/815876e7097fcd1e317ecfd867db807681f9cb1a#diff-c03ca828c6b8a7695f2cd7e52c316a3cR1188 back to u16 if you're so passionate 14:08 hmmmm you got all uptight over me pointing out that unsigned types should be preferred 14:08 Zeno` where? 14:08 Zeno` I said int 14:08 Zeno` int is signed 14:08 Zeno` you're confusing 2 different commits 14:09 hmmmm because they're both the same issue 14:09 Zeno` no, the previous commit was to fix a warning 14:10 Zeno` the commit you commented on is not related in any way and my previous commit did not influence *at all* my changing of the types of i 14:11 Zeno` but change them back to u16 if you like 14:11 Zeno` makes no difference to me 14:11 hmmmm you changed them for different reasons, but the underlying concern is that you both changed them to signed types 14:11 hmmmm that's all 14:11 hmmmm this is so asinine 14:11 Zeno` I changed them for different reasons 14:12 hmmmm that you did 14:14 Zeno` for (i = 0; i < 2; i++) { if (i == 0) {...} else if (i == 1) {...} } is crazier than what I did anyway heh 14:18 Zeno` or... ahem... does that construct make sense? 14:19 hmmmm you know you can be correct without needless pedantry 14:21 Zeno` Seriously, the change I made in the commit you're talking about from u16 to int was made without any thought whatsoever; it was just automatic 14:21 Zeno` change them back 14:31 luizrpgluiz hi 14:31 gregorycu hi 14:35 Zeno` hmm I think you should delete your comment 14:36 luizrpgluiz what could be the new version of minetest beyond the performance? 14:39 Zeno` performance 14:39 gregorycu Especially if you're on windows ;) 14:39 gregorycu Anyway, I can't repo that bug Zeno 14:40 gregorycu It's a MingW build 14:40 Zeno` gregorycu, better add a comment to the bug report then 14:40 gregorycu I'll see if I can repo with the same package 14:42 luizrpgluiz next year will leave the version 0.4.11? 14:50 gregorycu I can't repo with either 32 or 64 bit versions, I've update the issue 14:51 luizrpgluiz :) 14:59 hmmmm https://github.com/kwolekr/minetest/commit/7dbb3916cbc7f8650664db327ae35d72d3a872fb 15:02 luizrpgluiz but I say leave officially 15:03 Zeno` hmmmm, can that be merged since we're in feature freeze? 15:03 Zeno` It's not exactly a bug fix 15:03 hmmmm it can be if you frame it a certain way 15:05 Zeno` I guess it looks cool and all, but what about the current blockers? 15:05 hmmmm not aware of any blockers 15:05 Zeno` https://github.com/minetest/minetest/labels/blocker 15:07 hmmmm windows specific issue that i can't do anything about, font_size setting error we previously discussed and i thought we came to a conclusion on, bug there's a patch for which has a huge horrible bug i asked him to fix but he never did 15:10 luizrpgluiz it is a big mistake, it would be possible to fix it? 15:14 Zeno` hmmmm, you have a solution to #3? 15:15 hmmmm i guess we can cherry pick the commit, fix it, then push to upstream 15:15 Zeno` I guess if you can fix it you should 15:20 luizrpgluiz my dear developers, I'm just a player, I would have to study well the C and C ++ and LUA to help you fix many bugs in the game 15:21 luizrpgluiz I have many ideas for the game 15:24 gregorycu Everybody is an ideasman 15:25 gregorycu You play on windows? 15:25 hmmmm probably a discussion better for #minetest 15:25 kilbith ^ 15:26 gregorycu If you do play on windows, can you try and repo this bug: https://github.com/minetest/minetest/issues/1959 15:35 hmmmm realbadangel's fix doesn't even work 15:35 hmmmm i just tested it 15:35 Zeno` what fix 15:36 hmmmm for the wieldhand being fully bright in dark scenes 15:37 Zeno` I can't see the fix 15:37 Zeno` link? 15:37 hmmmm https://github.com/RealBadAngel/minetest/commit/57c6e96a4e23781fa190c6ee9925bbb660187204#diff-722710b5ecd9db42e7dd785dd517e186R919 15:38 Zeno` That's not a fix for the issue 15:39 hmmmm erm... it was referenced in the issue 15:40 Zeno` not by RBA 15:43 Zeno` I thought you had your own fix 15:59 Zeno` https://github.com/minetest/minetest/commit/7dbb3916cbc7f8650664db327ae35d72d3a872fb <--- WTH is this? 15:59 Zeno` there are no comments in the code to suggest it's temporary 16:00 Zeno` hmmmm ^^^ 16:12 gregorycu lol 16:13 gregorycu I just drastically increased the performance of the client in debug by replacing some vector with vector 16:13 gregorycu (Not even kidding) 16:13 gregorycu 2fps -> 30fps 16:15 Calinou where did you do that, which file/line? 16:16 gregorycu collision.cpp 16:16 gregorycu 231, 232, 233 16:18 Calinou does it work on release builds? 16:18 gregorycu About to find out 16:20 sfan5 gregorycu: bool shouldn't be that slow 16:20 celeron55 gregorycu: how did you figure out that optimization? 16:20 gregorycu Someone forgot to tell microsoft 16:20 celeron55 sfan5: it can be; vector is a special implementation 16:20 gregorycu I profiled 16:20 celeron55 what tool? 16:20 gregorycu Visual Studio Profiler 16:21 sfan5 hm 16:21 gregorycu In that function, there are 6 vectors 16:21 sfan5 celeron55: wouldn't it make more sense to use std::bitset 16:21 gregorycu However, most of that function was spend in std::vector::push_back 16:21 gregorycu Even though only 3 of the vectors were bool 16:21 Calinou the collision code is quite bad currently, I can easily believe it 16:21 Calinou the one who added nodeboxes worsened it 16:21 Calinou before it was reliable, you never got stuck anywhere 16:22 Calinou now, when you jump, you have 10 % chance to get stuck 16:22 celeron55 can someone else benchmark this change? 16:22 celeron55 2fps -> 30fps does not sound realistic though, that must be some kind of a specific situation 16:23 celeron55 like... walking on top of very complex nodebox thinggs 16:23 gregorycu It was in debug 16:23 gregorycu I've just profiled release 16:23 gregorycu One sec 16:25 celeron55 the speed efficiency of a vector in visual C++ and g++ might differ a lot 16:26 celeron55 kahrl: i think you made most of that code; comments? 8) 16:26 sfan5 hm 16:27 sfan5 why does it stay stuck at 61 fps even with wanted_fps=max_fps=999 and vsync=false 16:28 gregorycu There looks to be some improvement, but not much 16:28 gregorycu Not as much as was in debug, buy far 16:29 gregorycu I was running singleplayer, let me try server-client 16:29 sfan5 singleplayer is server-client 16:29 sfan5 just local 16:29 Calinou sfan5, fps_max 0 16:29 Calinou not max_fps 16:29 Calinou also you can use 0 to disable limit 16:30 Calinou and wanted_fps should always be 30 16:30 sfan5 *** Error in `./minetest': double free or corruption (!prev): 0x0000000002dabd60 *** 16:30 sfan5 i dun even 16:30 sfan5 I just ^C'd it 16:32 sfan5 ooh, 72 fps 16:32 celeron55 sfan5: fps_max? 16:33 sfan5 celeron55: nvm 16:34 sfan5 79 fps now 16:34 celeron55 also, calinou is very correct in that changing wanted_fps is optional 16:34 celeron55 it only affects the automatic view distance tuner 16:35 Calinou I used to use a big value like 100, and it severely decreased my FPS (I don't know why) 16:35 Calinou this was a while ago (2011-08), so the new autotuner may behave better 16:37 celeron55 a high fps in minetest won't feel very good on many systems though because some things tend to cause too large frametime spikes 16:37 celeron55 be happy if you happen to have one that doesn't suffer from that 16:38 gregorycu Well, I think there is some positive impact 16:38 gregorycu But it's not really conclusive in release 16:38 gregorycu In debug, there certainly is 16:38 gregorycu Massively 16:40 gregorycu Under Game:run, 75% of the time is spent in vector::push_back 16:40 gregorycu In debug 16:41 gregorycu This is Visual Studio 2013 16:41 gregorycu Anyway, I'm off to bed 16:42 celeron55 i think practically nobody is using MT in MSVC debug mode; not sure if this will yield any changes 16:42 celeron55 but good night 8) 16:52 sfan5 celeron55: i get about 4-7 more fps with the change 16:53 celeron55 for real? 16:53 celeron55 i don't even believe that, i guess i'll test it myself 16:57 celeron55 no difference 17:03 celeron55 yes, no measurable difference 17:04 sfan5 ¯\_(ツ)_/¯ 17:05 celeron55 measuring this would require calculating an average FPS over time and it would be some small portion of one, i guess 17:05 sfan5 why are we not using std::bitset? 17:05 celeron55 you can't make a dynamically sized one 17:06 celeron55 and it isn't optimized for speed either 17:06 sfan5 maybe we should write our own bitset implementation 17:07 sfan5 or maybe it's already fast enough currently 17:10 celeron55 what's the problem with vector or vector 17:11 celeron55 it doesn't even have a performance difference except for MSVC debug builds, apparently 17:22 shadowzone Gotta go 17:40 rubenwardy Can another core developer take a look at this? https://github.com/minetest/minetest/pull/1977 18:28 MinerDad I've created a pull request for inventory prediction (bug 1374): https://github.com/minetest/minetest/pull/1988 18:32 Calinou can it know if a move is rejected by server? does it revert them gracefully? 18:32 MinerDad yes 18:33 MinerDad there was already some client prediction, I just extended it to include more cases 18:54 VanessaE Zeno's attempt to fix plantlike did not work correctly. 18:55 VanessaE the size is right, but vertical positioning is broken again 18:55 VanessaE http://digitalaudioconcepts.com/vanessa/hobbies/minetest/screenshots/Screenshot%20-%2012182014%20-%2001%3a54%3a34%20PM.png\ 18:55 VanessaE http://digitalaudioconcepts.com/vanessa/hobbies/minetest/screenshots/Screenshot%20-%2012182014%20-%2001%3a54%3a34%20PM.png 18:55 Krock #1987 is quite interesting 18:55 ShadowBot https://github.com/minetest/minetest/issues/1987 -- Change TileSpec::frames to be std::vector not std::map by gregorycu 18:55 VanessaE that 45-degree plantlike part is not supposed to be embedded into the stalk of the tree, it's actually placed in that empty-ish space above it between the raillike parts :) 18:56 VanessaE do I need to open a new issue just to have it marked as a blocker or can someone just fix it? :) 18:57 VanessaE (zeno wanted me to test but I was unavailable at the time) 19:27 VanessaE https://github.com/minetest/minetest/issues/1989 19:39 Krock There's a data possible loss while converting lua_Number to char32_t in l-mapgen.cpp 404 19:40 Krock Could someone include that fix to an other one? 19:44 Megaf #1989 19:44 ShadowBot https://github.com/minetest/minetest/issues/1989 -- Plantlike visual_scale causes plants to be shifted downward 19:45 Megaf we should just rewrite minetest... 19:46 kaeza I'll let you start doing that 19:46 kaeza please, now, do it 19:46 VanessaE ... 19:57 Megaf kaeza: rightm, if I do, will you join me? 20:06 Bratmon Rewriting minetest should take what, the rest of this afternoon? 20:07 Bratmon Maybe the evening? 20:09 Megaf MInetest is quite simple actually 20:09 Megaf the engine itself can be very simple and all stuff could me added as mods 20:10 Megaf imagine an engine writen using this engine, https://en.wikipedia.org/wiki/Horde3D 20:13 Bratmon I'm not a graphics programmer. What exactly is the difference? 20:14 Megaf both Horde3D and Irrlicht are 3D engines that are suposed to be lightweight 20:14 Megaf but horde supports Lua and Python out of the box 20:14 Bratmon Okay. 20:16 Megaf VanessaE: what do you think? http://horde3d.org/features.html 20:18 Megaf well, minetest problem is not exactely the engine itself 20:18 Megaf well, not the 3D engine, Irrlicht is pretty capable 20:18 Megaf The problem is the way it begun and the way it is now 20:19 Megaf linux Winfows 9.x... 20:19 Megaf Windows* 20:20 Megaf If I had the knowhow to re write it I would 20:23 PilzAdam Megaf, it only supports OpenGL and is probably not suitable for old hardware 20:23 Megaf It isn't 20:23 Megaf I just read that now 20:23 Megaf and it doesn't seems to support EGL 20:24 Megaf I would rewrite minetest engine if I had the know how 20:49 VanessaE Megaf: we don't need a new engine. 20:50 VanessaE I won't say what I think we DO need because I'll undoubtedly stick my foot in my mou--keyboard. 20:50 Megaf no, we need the same minetest, but rewriten, with all dead/old code removed 20:56 VanessaE Megaf: the graphics stack upon which a game engine runs is also called an engine 20:56 celeron55 Megaf: if you had the know how, you would have much more interesting things to do 20:56 VanessaE e.g. irrlicht --> "The Irrlicht Engine is an open source high performance realtime 3D engine written in C++." 20:59 Megaf VanessaE: I like Irrlicht, but some of the problems that Minetest has it's not even Minetest fault, it's Irrlichts 21:00 Megaf celeron55: you mean, in real life or helping minetest ? :P 21:01 twoelk If I had the knowledge I might design myself a spacecraft, fly to the moon, laser myself some nice moon rock cubes and recreate some Minetest worlds in a 1node = 1meter scale. 21:02 * twoelk runs outside and stares at the moon 21:02 celeron55 lol 21:02 celeron55 it's only halfway rocket science 21:03 celeron55 Megaf: are you able to make a list of everything that you know is irrlicht's fault? 21:03 celeron55 because that would be interesting and useful 21:03 celeron55 (and i would also be able to correct you where you are wrong) 21:04 Megaf celeron55: Yep, I can list two issue right now 21:04 celeron55 i'm not interested in randomly picked issues, they aren't very constructive in planning for the far future 21:05 celeron55 a complete list would be 21:05 Megaf celeron55: so, EGL, the support is experimental on Irricht and it doesnt work on the Raspberry Pi 21:05 Megaf but fortunetaly people are making progress on Gallium 21:05 Megaf and copy and paste on Linux 21:06 Megaf Copy and paste. Proper font handling (not in a hacky way like sapier is doing) 21:06 hmmmm how is sapier's font handling hacky? 21:07 neoascetic can someone tell me in which file server pushes world state to client? 21:08 hmmmm server.cpp 21:08 Megaf hmmmm: he made an image for every letter 21:09 celeron55 have you ever heard of bitmap fonts? i don't understand why so many of them were added to the repo but yes, irrlicht is quite oldschool in not natively supporting others (MT supports vector fonts though) 21:22 Megaf celeron55: so why arent they used? 21:22 Megaf anyway, it worked with TTF 21:23 VanessaE the better question is, why do we still care about bitmapped fonts/non-truetype in this day and age? 21:23 VanessaE is there some random platform on which freetype can't be depended upon? 21:24 Megaf VanessaE: GNU/Linux, BSD, OS X, BeOS/Haiku, Windows all support TTF 21:24 VanessaE I'm with c55 on this one, the addition of those bitmapped fonts doesn't make a lot of sense to me either. 21:24 VanessaE Megaf: those are operating systems, I meant hardware platforms 21:40 RealBadAngel hi 21:40 VanessaE hi 21:41 roniz hi 21:46 RealBadAngel why https://github.com/minetest/minetest/commit/815876e7097fcd1e317ecfd867db807681f9cb1a ? 21:46 RealBadAngel this completely breaks the fixes to plantlike i made lately 21:47 roniz i am using 0.4.10 21:47 roniz how do I generate a v7 mapgen through the linux command line? 21:47 roniz because I only have ssh access to the server 21:48 VanessaE RealBadAngel: your fixes broke plantlike anyway... 21:48 RealBadAngel VanessaE, how exactly? 21:48 VanessaE RealBadAngel: the stuff we discussed in /msg ? 21:49 VanessaE or just #1978 anyway 21:49 ShadowBot https://github.com/minetest/minetest/issues/1978 -- plantlike drawtype visual_scale is way too small now 21:49 RealBadAngel its NOT to small 21:49 VanessaE it IS too small 21:49 RealBadAngel its correct (or rather was) 21:49 RealBadAngel before it was too big 21:49 VanessaE well was, until that commit that resulted in bug #1989 21:49 ShadowBot https://github.com/minetest/minetest/issues/1989 -- Plantlike visual_scale causes plants to be shifted downward 21:49 VanessaE no, you can't change the API like THAT 21:50 VanessaE that's backward incompatible 21:50 RealBadAngel ofc it wasnt backwards compatinble 21:50 RealBadAngel now you have same shit again 21:50 RealBadAngel and floating and improper scaled plants again 21:51 VanessaE yeah, I don't know why zeno committed this without giving me a chance to test it (he DID want me to test but I was afk) 21:51 RealBadAngel you just used to the bug and start called it normal behaviour dammit 21:51 VanessaE um, visual_scale = 2 making a plant twice as big is not a bug... 21:51 RealBadAngel go on, make now plant twice as big 21:52 RealBadAngel shit 21:52 VanessaE the only problem here is the vertical positioning of the plant. the scale is perfect. 21:53 RealBadAngel code doesnt care about visual scale other than 1.0 and above 21:54 RealBadAngel plant will be allways burried 21:54 RealBadAngel https://github.com/minetest/minetest/commit/815876e7097fcd1e317ecfd867db807681f9cb1a#diff-c03ca828c6b8a7695f2cd7e52c316a3cR1191 21:54 RealBadAngel this is fucking WRONG 21:55 RealBadAngel whole commit is just a mess 21:57 VanessaE I didn't write the patch but what I *do* know is the correct place to fix this is there at line 1191/1192 21:58 VanessaE those equations at 1174 and prior need to remain effectively the same or it fucks up the size of the image 21:58 RealBadAngel nvm, i will trash this code anyway. plantlike will be also converted to meshnode 21:58 VanessaE no, it's too late in the cycle 21:58 VanessaE we're in feature freeze now 21:59 VanessaE release is in less than a week 21:59 RealBadAngel i have made the conversion code month ago 21:59 RealBadAngel wont push it now, dont worry 21:59 RealBadAngel but its obsolete anyway 22:00 RealBadAngel what im angry at he broke the positioning of the plants 22:00 RealBadAngel without even asking bout it 22:01 VanessaE he asked. 22:01 VanessaE I wasn't here to test his fix 22:01 VanessaE you weren't here to comment 22:01 VanessaE idk why he proceeded though 22:01 RealBadAngel btw, funny to see dozens of new commits in feature freeze phase and see my commit blocked by PilzAdam 22:02 VanessaE also: 22:02 RealBadAngel which wasnt a feature 22:03 RealBadAngel VanessaE, yes? 22:03 VanessaE I'm comparing the code from before you changed it (when you broke the visual scale) to after zeno's last patch 22:03 RealBadAngel i havent broke it, i just set it to regular and correct scale 22:03 VanessaE you broke it. 22:03 RealBadAngel before it was broken 22:03 VanessaE the scale was fine before. 22:04 VanessaE the positioning is what was broke. 22:04 VanessaE anyway the first part up to line 1174 basically looks like a complete revert of your original patch 22:04 VanessaE (well 1174 as it exists today) 22:05 VanessaE my gut keeps telling me this could be solved by just adding BS / 2 to the equation at line 1192 and losing the conditional. 22:05 VanessaE but that's probably wrong (I haven't tried it) 22:05 VanessaE something like, vertices[i].Pos.Y -= BS/2 + BS/2 * (1 - f.visual_scale); 22:06 hmmmm https://github.com/blog/1938-git-client-vulnerability-announced 22:06 PilzAdam RealBadAngel, your judgement of feature / bugfix, backwards compatibility and broken / non-broken lets me wonder if you are still a suitable for being a core dev 22:08 RealBadAngel PilzAdam, you are ALWAYS against what im doing, and that pisses me off 22:08 VanessaE RealBadAngel: see above. What effect would that change have? 22:08 VanessaE enough of the banter, let's just fucking FIX THIS RIGHT? PLEASE. 22:08 RealBadAngel VanessaE, i will try it 22:09 RealBadAngel gimme a few minutes to test it 22:10 RealBadAngel PilzAdam, and fix that damn screwdriver rewrite, its broken 22:10 VanessaE screw it, I'll just try it myself. 22:10 PilzAdam RealBadAngel, I'm not always against what you are doing; e.g. the meshnode drawtype is cool 22:11 PilzAdam btw, RealBadAngel, https://github.com/minetest/minetest_game/pull/379 22:11 RealBadAngel PilzAdam, you have asked me whats wrong with rewrite then pissed on my opinion 22:12 PilzAdam yes, I asked what is wrong with the rewrite, not which bugs already existed before the rewrite 22:12 PilzAdam you seem to still not understand that 22:12 RealBadAngel PilzAdam, dont be so strict 22:13 RealBadAngel that was a good time to take care of it and you know that 22:14 PilzAdam it wasn't; the "rewrite" just changed the way the player interacts with it; not how it internally works 22:14 PilzAdam anyway, thats not the point now 22:14 RealBadAngel oh, ok 22:15 RealBadAngel about plantlike, imho it should be aware of it 22:15 PilzAdam you generally seem to have a quite different judgement of things than I do 22:16 RealBadAngel PilzAdam, quite normal, im not a copy of you, nobody is ;) 22:17 RealBadAngel in general we should consider a flag for screwdriver 22:17 PilzAdam for example, you think breaking backwards compatibility is reasonable if you think the current way is "broken" 22:17 PilzAdam (e.g. plantlike) 22:17 RealBadAngel yes, sometimes i do 22:17 PilzAdam it's not 22:18 PilzAdam it's even worse when notifying nobody about it 22:18 RealBadAngel if plant is floating or being burried (depending on the scale) and API says it should be positioned on the ground... 22:18 RealBadAngel wheres the bug? 22:18 PilzAdam e.g. nobody knew that you intentionally broke the API, hence Zeno's "fix" 22:19 RealBadAngel everybody in the channel knew 22:19 RealBadAngel you was away propably 22:19 PilzAdam not enough people, it seems 22:19 RealBadAngel that change was discussed 22:19 PilzAdam with whom? 22:19 RealBadAngel im not pushing anything lately without discussion 22:20 RealBadAngel see the commit and pictures 22:20 RealBadAngel and the issue which was opened for that 22:21 PilzAdam API changes previously even had forum posts: https://forum.minetest.net/viewtopic.php?f=18&t=6808 22:21 PilzAdam "everyone in the channel" is not enough 22:21 PilzAdam it is a project goal to have a stable API between releases; you violated that goal 22:23 RealBadAngel https://github.com/minetest/minetest/issues/1747 22:24 RealBadAngel do you think that was a normal behaviour? and that was api claimed? 22:25 PilzAdam you broke VanessaE's mod 22:25 gregorycu What is the MeshUpdateThread? 22:25 RealBadAngel bug being not fixed for too long became a feature 22:26 RealBadAngel but thats not an excuse to not fix it 22:26 VanessaE there was nothing wrong with the visual_scale size range 22:26 PilzAdam how is making nodes 10 times bigger a fix for floating plants? 22:26 VanessaE only the Y position was bad. 22:26 RealBadAngel at which point something was 10x bigger?? 22:26 PilzAdam see VanessaE's original issue 22:27 VanessaE RealBadAngel: that was an entity with mesh drawtype! 22:27 RealBadAngel PilzAdam, i think you messed up threads 22:28 PilzAdam #1978 22:28 ShadowBot https://github.com/minetest/minetest/issues/1978 -- plantlike drawtype visual_scale is way too small now 22:28 PilzAdam s/bigger/smaller 22:29 PilzAdam there are other things that make me question your judgment of things, too (besides your views on compatibility) 22:29 RealBadAngel VanessaE, thx to my fixes plants were perfectly the same size as other "pixels" around 22:30 PilzAdam there were 2 incidences of people bisecting a problem down to one of your commits; and you simply didn't accept it and said it was caused by someone else 22:30 RealBadAngel you just got used to buggy behaviour and called it NORMAL 22:31 PilzAdam and then there is this screwdriver thing, where you were simply unable to judge the contents of a PR and constantly pointing to an unrelated / out of scope bug 22:31 RealBadAngel PilzAdam, afaik last time some1 was bisecting someting i fixed that 22:32 PilzAdam RealBadAngel, one of it was me trying to find the cause for #1705 22:32 ShadowBot https://github.com/minetest/minetest/issues/1705 -- Regression in smooth day-night transition 22:32 gregorycu You two should stop fighting :P 22:32 PilzAdam after bisecting, you tried to tell me that this wasn't a regression 22:32 RealBadAngel PilzAdam, this is also fixed 22:32 gregorycu And someone should answer my question 22:32 PilzAdam and that your commit didn't cause it 22:32 PilzAdam RealBadAngel, finally, after hours of discussing 22:33 RealBadAngel PilzAdam, i was hoping for another solution 22:33 RealBadAngel sadly in both cases i was forced to get back to old code 22:33 PilzAdam then you should have said that, instead of saying it's not caused by you 22:33 RealBadAngel i did 22:34 PilzAdam do you really want me to dig up the logs? 22:34 RealBadAngel but youre lately not around :P 22:34 RealBadAngel if i didnt, i wouldnt be fixing that 22:34 PilzAdam I'm not saying that you are "not good", or "bad"; I'm just saying that your way of reasoning seems to be quite different to mine 22:35 PilzAdam that makes us quite incompatible as discussion participants 22:35 PilzAdam and I don't like that 22:35 RealBadAngel do you expect me to like that on the other hand? 22:36 RealBadAngel i always find you in the opposite corner of the ring 22:36 PilzAdam it's not about oppinions (I think) 22:37 neoascetic why server.cpp deals with playersao while ServerEnvironment deals with player object? 22:39 PilzAdam RealBadAngel, also, because of bias, I think that my "way of thinking" is better, so I'd like to see your right to vote on PR's removed 22:40 VanessaE while you guys are bitching at each other, I'm actually trying to fix this the right way 22:40 VanessaE http://pastebin.com/9XSiq5t6 22:40 VanessaE there. 22:41 VanessaE that works fine for anything that's visual_scale = 1.0 or larger. 22:41 VanessaE no change to anywhere else in the code. 22:41 gregorycu 8% of the render loop is checking settings, just saying 22:41 RealBadAngel VanessaE, please dont reinvent the wheel: https://github.com/minetest/minetest/pull/1871/files#diff-c03ca828c6b8a7695f2cd7e52c316a3cR1183 22:42 VanessaE I don't have anything handy that has a scale of less than 1.0, can someone please check? 22:42 RealBadAngel this code doesnt use any mind tricks, just the same visual scale for everything 22:42 VanessaE RealBadAngel: look, I don't know enough C++ to even begin to touch other parts of the engine, this little piece I can at least understand. 22:42 VanessaE this works. 22:43 RealBadAngel ofc to some point 22:43 VanessaE screw it, I'll just test the small stuff too 22:43 RealBadAngel main problem of the drawtypes is they were written by different folks 22:43 hmmmm breaking compatibility is bad 22:43 PilzAdam RealBadAngel, to put it in other words: I don't trust you to judge PRs the same I trust the other devs 22:43 hmmmm and, yeah, if a bug stays around long enough, it becomes the expected version 22:44 hmmmm to 'fix' this, have an option to use the fixed version 22:45 PilzAdam RealBadAngel, is my view on this understandable or do you think I'm just an ass? 22:46 VanessaE opk, confirmed 22:46 RealBadAngel youre stubborn the same way as i do :) 22:46 VanessaE my fix works. 22:46 VanessaE PR incoming in a few minutes. 22:46 VanessaE or can someone just take that paste and use it please? 22:46 VanessaE save me a little trouble :) 22:48 RealBadAngel hmm, but such bugs are pain in the ass, like the positioning in plants, or torchlike being upside down for param2 = 0 22:48 RealBadAngel hmmmm, ^ 22:48 hmmmm boo fucking hoo 22:48 RealBadAngel keeping compability means extra code in rendering loops 22:48 RealBadAngel which is not wise 22:49 hmmmm you can remove the extra compatibility code when 0.5.0 comes out 22:50 VanessaE https://github.com/minetest/minetest/pull/1991 22:50 VanessaE there 22:50 VanessaE assuming I did it right, this works for me 22:50 RealBadAngel hmmmm, thats why im not pushing my recent changes into current tree, atm i trying to convert as much as i can to native irrlicht meshes 22:51 RealBadAngel this will speed things up and solve the weird issues like with plants 22:51 RealBadAngel one model for an object and let irrlicht rotate it or scale 22:55 VanessaE um, hello? 22:56 VanessaE are you guys gonna bitch at each other or is someone gonna take a look at a one-line pull to fix this issue, and at least tell me if I should go jump in a lake? 22:56 RealBadAngel VanessaE, hrm 22:56 RealBadAngel arent those two lines mathematically the same? 22:56 VanessaE s/at a/at that/ 22:56 VanessaE RealBadAngel: obviously not, or it wouldn't have fixed it 22:56 RealBadAngel https://github.com/minetest/minetest/pull/1991/files#diff-c03ca828c6b8a7695f2cd7e52c316a3cL1192 22:57 VanessaE besides, 1 - X is rarely equivalent to X - 1 :-) 22:57 VanessaE (unless X happens to equal 1) 22:57 PilzAdam RealBadAngel, they are 22:57 PilzAdam (mathematically equivalent) 22:58 jin_xi so, lake it is 23:00 VanessaE well somehow it works. 23:01 VanessaE maybe just because of the lack of the conditional. 23:08 RealBadAngel sorry, seems like i had problems with my connection 23:12 * VanessaE wanders off 23:12 VanessaE (it'll take me a few minutes to get to the nearest lake to make the aforementioned jump) 23:13 jin_xi why? did you test again? was it a ghost fix? 23:13 RealBadAngel jin_xi, zeno's fix broken my fix then VE fixed it again 23:14 VanessaE jin_xi: no, it works, but the real fix was just getting rid of that conditional; my equation is mathematically the same as the original. 23:14 RealBadAngel VanessaE, not the original 23:14 RealBadAngel original was fucked up, youre at the point of my fix 23:15 VanessaE well same as what's already there. 23:15 RealBadAngel but with opposite signs just lol ;) 23:16 VanessaE well with that little fix, now plantlike behaves exactly like it should 23:16 VanessaE corret sizing, correct positioning 23:16 VanessaE or so it appears. 23:18 hmmmm so 23:18 hmmmm I really don't like https://github.com/minetest/minetest/commit/7dbb3916cbc7f8650664db327ae35d72d3a872fb 23:19 hmmmm i can remove that if minetest_game and minimal have a subset of paramat's biomes 23:19 hmmmm thoughts? 23:20 VanessaE this would affect mgv5 and mgv7 only I presume? 23:20 hmmmm yep 23:21 VanessaE yeah, seems reasonable to me, but expect complaints about lack of beaches. 23:21 hmmmm don't care 23:21 VanessaE (but, better that ^^^^ than all-stone) 23:21 hmmmm it'll be something different from the usual complaint 23:22 exio4 the whole map is just grass! 23:23 VanessaE then again the aforementioned subset would likely be more than just grass 23:33 RealBadAngel VanessaE, so im ok with the fix by now 23:34 VanessaE *GASP* 23:38 RealBadAngel before 0.4.11 i will fix just that issue with wielded being lit when shaders on 23:40 RealBadAngel the other code is not yet ready to go into stable, main change (moving highlighted and cracks to separate scene node) has to be widely tested before merging 23:40 RealBadAngel after that is done i can work on other things 23:41 RealBadAngel point is i need to have clean mapblock mesh, to turn it into mesh with tangents 23:43 paramat hmmmm i could submit a simplified version of my v5/v7 biomes branch if you want (to be replaced by the full system in 0.4.12): plains with grasses/flowers (no trees perhaps), simple beaches and sand underwater 23:43 hmmmm sure 23:45 paramat okay 23:51 paramat since schems by lua table is fixed i might also be able to whip up a quick appletree with randomised leaves and apples 23:54 paramat (but then again maybe not, i'd rather do the tree schems right or not at all ... okay just plains will do)