Time Nick Message 00:02 RealBadAngel kahrl, ive changed that in pr 00:04 kahrl well 00:04 kahrl with an increased angle you may just have to move farther away to get the problem again 00:05 kahrl but I can't load and render enough mapblocks to get the problem (with *1.1) before melting my computer 00:05 kahrl so I guess it's okay 00:08 kahrl (I'd like to know where the mathematical error in the existing function was) 00:08 RealBadAngel its over 20 yrs since my last math classes so i wont dig it ;) 00:26 RealBadAngel kahrl, i think i know whats wrong 00:27 RealBadAngel // Maximum radius of a block. The magic number is 00:27 RealBadAngel // sqrt(3.0) / 2.0 in literal form. 00:27 RealBadAngel f32 block_max_radius = 0.866025403784 * MAP_BLOCKSIZE * BS; 00:27 RealBadAngel what sqrt(3) is doing here? 00:28 RealBadAngel diagonal of a square is a * sqrt(2) 00:31 kahrl we have a cube though, not a square 00:31 RealBadAngel ooops 00:32 RealBadAngel i told i shouldnt dig it ;) 00:32 kahrl :D 05:20 paramat now merging game#839 05:20 ShadowBot https://github.com/minetest/minetest_game/issues/839 -- Consolidate ABMs by paramat 05:29 paramat merged 07:47 RealBadAngel https://github.com/minetest/minetest/pull/3659#issuecomment-183843567 07:47 RealBadAngel kahrl, do you agree with that? 08:01 RealBadAngel #3707 08:01 ShadowBot https://github.com/minetest/minetest/issues/3707 -- Use proper variable types for uniform sampler layers by RealBadAngel 08:05 RealBadAngel hmmmm, thx for noticing above, that fixed bug i couldnt trace when coding renderer 08:09 RealBadAngel nore, please reconsider increasing camera_offset greatly, https://github.com/minetest/minetest/issues/3694#issuecomment-183536991 08:10 RealBadAngel nore, the issue you described with nodeboxes was caused by rendering precision AND the depth buffer 08:11 RealBadAngel from what i can see it should be possible to set offset at least 1k or even more 08:12 RealBadAngel with BS set to 1.0 (that affects also rendering precision) glitches are barely visible ~10 k distance 08:59 est31 gonna push #3707 in 10 mins 08:59 ShadowBot https://github.com/minetest/minetest/issues/3707 -- Use proper variable types for uniform sampler layers by RealBadAngel 08:59 est31 its small Pr and pretty clear that the stuff it does makes sense 09:05 RealBadAngel est31, what bout 3705? 09:08 est31 last time I saw 3705 it set the angle far too wide 09:08 RealBadAngel changed that 09:08 est31 10 percent... 09:08 est31 well seems okay for me 09:09 RealBadAngel well 09:09 RealBadAngel with an increased angle you may just have to move farther away to get the problem again 09:09 RealBadAngel but I can't load and render enough mapblocks to get the problem (with *1.1) before melting my computer 09:09 RealBadAngel so I guess it's okay 09:09 est31 whats the actual issue here? 09:10 RealBadAngel blocks are disapearing in the corners, because of occlusion 09:10 RealBadAngel look at screenshot in PR 09:11 RealBadAngel visible at certain angles and heights 09:13 est31 ah I see 09:13 est31 hrmm could use a helpful comment 09:13 est31 I'll add a proposed comment on github 09:13 RealBadAngel ok 09:19 RealBadAngel est31, any comments on code #3695? 09:19 ShadowBot https://github.com/minetest/minetest/issues/3695 -- Use vertices with tangents only when its needed. by RealBadAngel 09:36 est31 RealBadAngel, commented on the PR 09:39 blunaxela For code style, do you all use a checker like the kernel checkpatch.pl, or would that be too strict? 09:40 est31 no, we review it manually 09:40 est31 but checkpatch.pl would be an idea 09:41 est31 but its written for C anyway 09:41 est31 so needs changes 09:41 est31 probably not worth the effort 09:41 est31 5.9k lines 09:42 blunaxela yeah, i assume it might choke on c++ a bit, oh, and it's quite large... 09:54 blunaxela interesting, i need to figure out how to traverse the entire tree still, but /src/*.cpp only came out with trailing whitespace 09:58 est31 there should be almost no file anymore 10:04 jhcole does anyone know how to remove a craft recipe 10:05 jhcole I'd like to use the trapdoor recipe to craft something else instead 10:25 blunaxela jhcole: (i'm very new to minetes) have you looked under /games/minetest/mods/doors/init.lua 10:25 blunaxela that's where the recipe seems to get registered 10:28 jhcole yes, i could just change it there. The down side is that it makes my minetest_game less like upstream 10:28 jhcole i was hoping to do it in the mod where i'm using the recipe to keep the relevant code all in one place 10:32 jhcole i did notice the new trapdoor_steel recipe uses a different pattern than the trapdoor recipe, so i will probably change trapdoor to match the pattern of trapdoor steel, and maybe submit a pull request 10:33 RealBadAngel est31, what for adding reserve(n) if push_back does the reallocation automatically? 10:34 est31 the reallocation is done automatically yes 10:34 est31 but doing it with reserve is faster 10:34 RealBadAngel i see 10:35 est31 if push_back does reallocation twice or more often 10:36 blunaxela jhcole: ah, so the current trapdoor to be 2x2, and yours would be 2x3 10:36 blunaxela jhcole: what is your new recipe out of curiosity? 10:38 jhcole blunaxela: i think the current trapdoor is 2x3, i'm suggesting making it 2x2 to match trapdoor_steel 10:39 jhcole blunaxela: and also only output one trapdoor instead of 2 10:41 Obani yo 10:49 jhcole blunaxela: i just reread what you wrote and think i completely misunderstood. My new recipe would be fences for all the different types of wood 10:49 RealBadAngel est31, vertex_count is number of vertices already stored, you propably meant nr of vertices im going to add, right? 10:51 est31 yes 10:51 est31 and the current size 10:51 est31 add them together 10:52 RealBadAngel so reserve(current + new) yes? 10:52 RealBadAngel do the same to p->indices? 10:52 est31 yes 10:55 RealBadAngel pushed changes, check it now 11:06 est31 RealBadAngel, got my +1 15:43 xeranas Hi, can I add callback function on node for "player inside node" event?. From docs 'hacky' flags like "post_effect_color", "damage_per_second" suggest that there arent. Or I'm wrong? 15:44 kaadmy afaik you have to check manually 15:46 xeranas so every mod author adds theyr own on_step where they checks if they standing on right node? Even if "right" nodes could be few nodes in whole chunk? 16:05 xeranas hmm on_step is for entities only 16:05 xeranas kaadmy: so how people checks if player steps on certain node (not entity)? 16:11 xeranas I talk about walkable nodes right now 17:03 nrzkt sfan5, nore can you look at #3708 please ? and vote 17:03 ShadowBot https://github.com/minetest/minetest/issues/3708 -- Player::accelerateHorizontal/Vertical should be member of LocalPlayer by nerzhul 17:04 sfan5 nrzkt: what about d_wanted, can that be moved too? 17:04 nrzkt i didn't look at this variable atm 17:04 nrzkt wait a minute 17:05 nrzkt d_wanted is a local function variable 17:05 nrzkt we don't need to move it 17:06 sfan5 oops 17:06 sfan5 didn't see the declaration 17:06 nrzkt no problem :) 17:06 sfan5 nrzkt: you have my approval for 3708 17:07 nrzkt thanks for the review 17:18 Hijiri xeranas: I think the mesecons pressure plate uses an abm or node timer and checks if a player is on top 17:19 Hijiri you could also use register_globalstep 17:20 xeranas Hijiri: does not look eficient methods, but seems its all what we have. Thanks! 18:37 everamzah !tell RealBadAngel the halo node highlighting goes opaque when there's two players? 18:37 ShadowBot everamzah: O.K. 18:38 everamzah oops http://i.imgur.com/h3zXr7f.png 18:40 kaadmy is the node hilighting work for multiple players now?!?! 18:40 kaadmy does* 18:44 everamzah kaadmy: oh it's not supposed to? 18:47 RealBadAngel funny, what the hell transparency of the material have to do with number of players online? 18:47 everamzah no clue, me 18:47 everamzah dunno why, just that it seems to be the case 18:47 Calinou I guess it's like that water transparency issue? 18:47 Calinou where water becomes opaque in some MapBlocks, but only in multiplayer 18:50 RealBadAngel lol 18:51 RealBadAngel that will be hard one to track i afraid 18:56 Fixer happy birthday, RealBadAngel 18:57 RealBadAngel thx :) 19:06 RealBadAngel Calinou, indeed both work together 19:30 RealBadAngel can anybody else vote for #3695? nrzkt, kahrl? 19:30 ShadowBot https://github.com/minetest/minetest/issues/3695 -- Use vertices with tangents only when its needed. by RealBadAngel 19:41 RealBadAngel Calinou, that really "works". and it is strange. open 2 clients, look at water. disconnect with one. water goes transparent. connect -> opaque 19:42 RealBadAngel simply connecting changes the water immediately 19:42 RealBadAngel halo suffers the same effect 19:44 RealBadAngel ive made a small pool at y = 400, and watching with pause on one client, the second player is standing at ground level, so far away 19:52 RealBadAngel i do have an idea 19:53 RealBadAngel those are materials changing somehow in the middle of rendering 19:53 RealBadAngel so, what changes when player joins? 19:54 RealBadAngel player nametag is being displayed 19:55 Fixer RealBadAngel, lol 19:56 Fixer RealBadAngel, may you fix all the bugz and not loose your insanity till the next birthday 20:01 RealBadAngel voild 20:01 RealBadAngel voila i mean 20:02 RealBadAngel ive disabled the bug 20:02 RealBadAngel content_cao.cpp line 978 is the source of transparency bug 20:03 RealBadAngel both for water and halo 20:05 Fixer interesting 20:05 Fixer i can test, if needed 20:05 RealBadAngel sure 20:06 RealBadAngel comment out all lines starting with m_textnode 20:06 RealBadAngel 978 - 986 20:07 Fixer RealBadAngel, transperency bug = that thing with opaque water and with windows? 20:07 RealBadAngel idk if with windows, but water and halo for sure 20:12 RealBadAngel wonder whats in usin text node is wrong 20:13 Fixer compiling 20:45 Fixer RealBadAngel, looks like it fixed the problem for me, looking more 20:59 Fixer RealBadAngel, interesting, no bug for me 20:59 Fixer RealBadAngel, bug is disabled 21:11 RealBadAngel Fixer, yeah, now have to find out whats wrong with text scene node... 21:15 Fixer RealBadAngel, yes, looks like water squares opacity, water opacity and windows opacity are fixed, however, other blinking bugs in complex liquids remain, like this https://i.imgur.com/Ucc5cXD.jpg (but it is very good start) 21:15 Fixer RealBadAngel, and damn it is slow without 3695 21:16 RealBadAngel the "other" bug is not related 21:16 RealBadAngel this is due to zbuffer sorting 21:16 Fixer RealBadAngel, yeah, i noticed they are different 21:16 RealBadAngel transparent stuff requires separate render pass 21:16 RealBadAngel will be done in the future 21:16 Fixer RealBadAngel, your code change fixed a LOT of blinking 21:17 Fixer checked on 2 servers 21:17 Fixer RealBadAngel, so very nice start 21:17 Fixer i have hope 21:17 Fixer RealBadAngel, player report another interesting bug to me, with glass and transperency 21:17 RealBadAngel most important in bugs elimination is to know the reason 21:17 RealBadAngel we do now 21:18 RealBadAngel bad news is that it looks like irr bug 21:19 RealBadAngel on the other hand i planned to remove nametags from the scene anyway 21:19 RealBadAngel they cannot stay there in current form 21:19 RealBadAngel have to be moved to HUD 21:20 Fixer RealBadAngel, have you seen smth like that? https://imgur.com/a/myQjU 21:20 RealBadAngel thats because of postprocessing. i cant shade texts with bloom dof etc 21:20 Fixer it is obsidian glass slabs 21:21 Fixer from one side they are ok, from another, they are transperent 21:21 Fixer they are from moreblock 21:21 RealBadAngel seems legit 21:21 RealBadAngel its glasslike drawtype 21:22 RealBadAngel glasslike framed should be used for such things 21:22 * Fixer pokes Calinou 21:23 RealBadAngel about nametags luckily i wrote waypoints code 21:24 RealBadAngel i will just move nametags into std::vector in hud and display them all same way as waypoints 21:25 RealBadAngel im not going to dig in irrlicht while i have to move them anyway, just havent thought i have to do that right now 21:26 RealBadAngel lol 21:26 RealBadAngel Fixer, you poked him too hard ;) 21:26 Fixer lol 21:28 Fixer RealBadAngel, so transperency could be fixed by mod maker? 21:30 RealBadAngel Fixer, i think its textured wrong way 21:30 RealBadAngel will check it later, but for sure you wont have such problems while using framed 21:31 RealBadAngel framed were designed for such things 21:37 Fixer !tell Calinou https://github.com/minetest-mods/moreblocks/issues/32 21:37 ShadowBot Fixer: O.K. 22:53 Hijiri Is it true that non-fleshy damage types don't have an effect on players, even if their armor group is set? 22:54 Hijiri I know that immortal armor group has no effect 22:58 Hijiri actually I just tested it, it does have an effect 23:02 Hijiri actually no 23:02 Hijiri that was just the attacking client 23:18 jhcole i've compiled the latest minetest server using DRUN_IN_PLACE=1 so i can do some tests before installing globally, but contrary to the dev wiki it is loading mods from '~/.minetest/mods/' instead of 'minetest-install-directory/mods/'. Does anyone know how to fix this? 23:19 jhcole Or how to run two separate instances of minetestserver, each with there own separate mod directories? 23:26 RealBadAngel if you compile in place theres no way it can use global mods 23:26 RealBadAngel propably you are running wrong bin? 23:26 Hijiri Is it possible to override how punch callbacks are called during a punch? 23:27 Hijiri I want to calculate the damage specially and then pass that to the existing callbacks 23:27 Hijiri if I overwrite minetest.register_on_punchplayer it won't change the behavior of existing callbacks 23:30 jhcole RealBadAngel: im running with ./bin/minetestserver in the git repo where i compiled 23:32 RealBadAngel jhcole, youre compiling the repos? better clone fresh sources, move them to destination folder and compile again 23:32 RealBadAngel propably you have there messed make files 23:35 jhcole RealBadAngel: perhaps, i'll try that, but i did pull the latest from the master branch, and git diff is null 23:36 RealBadAngel have you ever compiled game in this directory before? 23:36 RealBadAngel with different make options? 23:37 jhcole yes i have, I reran cmake though, is that not enough? 23:45 jhcole hmm, i'm seeing a warning from cmake 'Manually-specified variables were not used by the project: RUN_IN_PLACE' 23:45 jhcole oops, i had a lowercase l in PlACE 23:51 jhcole RealBadAngel: thanks for you're help, its working now :) 23:52 RealBadAngel if you want to clean it, you shall delete cmake_cache.txt 23:53 RealBadAngel CMakeCache.txt actually 23:57 Hijiri Is it okay to access core.registered_on_punchplayers 23:57 Hijiri in a mod 23:57 Hijiri or is it meant to be internal