Time Nick Message 01:33 paramat ugh i'm regretting making biomes, ores and decos relative to water level, it makes everything so complex and wasn't essential. looks like hmmmm was right in refusing that years ago 02:34 sofar it doesn't make sense in a geologic fashion either XD 02:40 paramat i'll probably revert those changes soon 02:40 paramat anyway, this should have been done a long time ago #6422 02:40 ShadowBot https://github.com/minetest/minetest/issues/6422 -- Generate biomes: Recalculate biome at biome lower limit by paramat 02:40 paramat separate issue 03:05 paramat after a little more testng i'll merge that as it is fairly trivial, obvious and a bugfix 03:59 paramat will merge that in a moment 04:05 paramat oh erm, after i update docs 04:12 paramat ok merging 04:18 paramat done 06:19 paramat sofar one for you #5342 06:19 ShadowBot https://github.com/minetest/minetest/issues/5342 -- Connected nodeboxes don't support facedir (anymore?) 16:23 Krock merging trivial rename PRs in ~15 minutes: #6421 #6420 16:23 ShadowBot https://github.com/minetest/minetest/issues/6421 -- Rename "itemstring" to "item name" (which is actually used) in tooltip setting by DTA7 16:23 ShadowBot https://github.com/minetest/minetest/issues/6420 -- fix addModsFormConfig typo by hashcacher 17:36 paramat game#1813 is holding up other work now, please could mtg devs consider it? 17:36 ShadowBot https://github.com/minetest/minetest_game/issues/1813 -- Ores (non-mgv6 mapgens): New distributions by paramat 17:37 paramat next i'd like to do this #6216 i hope people are ok with this? 17:37 ShadowBot https://github.com/minetest/minetest/issues/6216 -- Add player eye height as player object property 17:43 paramat closed #6423 *sigh* 17:43 ShadowBot https://github.com/minetest/minetest/issues/6423 -- Network protocol documentation by zaoqi 17:44 Krock no, paramat! why did you close that? it was my best favourite pull, ever! ;) /s 17:45 thePalindrome Huh 17:45 Krock I thought we had a eye position setter function somewhere 17:46 Krock apparently it's just an offset :/ 17:49 Krock hello numzero. Testing the recent commits right now. Well done, the crack animation is now scaled correctly 17:51 Krock which looks quite funny with a scale of > 1 :D 17:54 numzero :D thanks 17:55 Krock krock-works.16mb.com/u/xmas_blinky.mp4 17:55 Krock ^ scale 1 / 10 18:12 numzero seems I need to add some support for non-power-of-2 scales 18:15 numzero Krock: test now, should work 18:16 numzero still, scale should be power of 2 18:16 numzero textures won’t match at mapblock boundaries otherwise 18:26 Krock given a 256x256 texture and a scale of 1/16 should spread it perfectly on the mapblock, using 16x16 pixels per node, which it only barely does: https://i.imgur.com/x4Ame6Z.png 18:27 Krock but the problem is gone when using a scale of 1/8, so using 32 pixels per node 18:36 numzero works for me, at least with autoscaling 18:36 Krock hmm 18:39 numzero found 18:48 numzero oh well, that’s due to F1000 format used for serialization 18:48 numzero changed scale → 1/scale 18:48 numzero and F1000→U8 18:49 numzero pushing soon 18:50 numzero pushed (not tested, not even compiled: it recompiles 189 files) 18:52 Krock 189? that's crazy 18:58 numzero that’s common after header change, esp. after rebase (there was an 1-line change in networkprotocol.h...) 19:00 numzero also I see crack size is still wrong sometimes; it tends to tile it on normal nodes (at least in the version I tested) 19:28 Fixer https://github.com/Thomas--S 19:28 Fixer https://github.com/minetest/minetest/pull/5919#issuecomment-329986736 19:35 thePalindrome hmm 19:40 numzero Krock: now should work. Crack size may be incorrect sometimes, however, if autoscale is in effect 19:45 numzero really, it’s better not to use the same texture with different scale 19:45 Krock ah yeah, now it looks nice :) 19:46 Krock what kind of problems would this cause? 19:46 numzero 16x16 crack grid 19:47 numzero on a node 19:47 numzero smth like that 19:47 Krock ah I see. 19:51 numzero trying to fix... 19:51 Krock impressive enough https://i.imgur.com/tSxupJX.png 19:58 thePalindrome Is it just me or does v7 have more floating bits? 19:59 sfan5 pretty sure that's intended 19:59 Calinou it does have an high concentration of floating islands, compared to v6 which doesn't have any (it uses 3D noise) 19:59 Calinou 2D noise* 19:59 Calinou v7 is 3D noise, like v5, v6 is 2D noise (faster, but overhangs are impossible, just like early Minecraft versions) 20:00 celeron55 that's not 100% correct; v6 makes them sometimes when the large-scale carving thing happens to make one 20:00 celeron55 but it's rare 20:00 celeron55 or was it disabled at some point? 20:01 celeron55 maybe it was 20:05 Calinou it would be nice if v6 used Simplex noise, 2D Simplex noise does not have patent concerns, only 3D and more 20:07 Krock One can not simply switch the map generation algorithm without making borders of old worlds ugly. 20:10 Hijiri rubenwardy: how do I display a warning? 20:11 Krock ^ warningstream ? 20:11 Hijiri where do I obtain a warningstream? 20:11 Hijiri Does it need to be passed as a parameter into my function? 20:11 sfan5 it's global 20:11 Hijiri ah, ok 20:20 Fixer thePalindrome: it does from time to time 20:34 Krock merging #6424 in 5 mins 20:34 ShadowBot https://github.com/minetest/minetest/issues/6424 -- Fix Rotate Node Placement by tenplus1 20:34 sfan5 when's 0.5 release planned actually? 20:35 nerzhul not really planned 20:35 nerzhul maybe end 2017 or beginning 2018, we should release a 0.4.17 in november with the bugfixes 20:35 nerzhul a backport campain should be done at a point 20:35 sfan5 >end '17 20:35 sfan5 ;_; 20:35 sfan5 nerzhul, btw i have updated the list of fixes that should be backported you created on gh 20:35 nerzhul there are many other things to add i think, mod channels and enhanced CSM :p 20:36 nerzhul sfan5, nice, ty 20:36 Krock heh, so it's going to be an xmas present like 0.4.15? ^^ 20:37 nerzhul i don't know if it's possible, but at a point we should have a feature freeze, then we need to identify remaining blocking/breaking points to change before release 20:37 nerzhul i need also finish to study using asio to have modern network subsystem , but it can be delayed 20:37 sfan5 async io is not necessary to fix networking 20:38 nerzhul not it's exact but it permits to have some facilities for it :) 20:38 nerzhul with Krock we really make it cleaner, it'sn ot yet finished but it's far more better to maintain 20:39 nerzhul made* 20:40 nerzhul sfan5, i think a feature we should introduce is to use zstd instead of zlib for new mapblocks 20:40 nerzhul and network comprssion 20:40 sfan5 you mean brotli surely 20:40 nerzhul i tend to prefer zstd and lz4 :p 20:40 sfan5 (zstd is faster but means bigger maps) 20:40 nerzhul how many percent difference between the two comprssion algos ? 20:41 sfan5 idk 20:41 nerzhul in terms of size i mean :p 20:41 sfan5 the exact measurements are in some issue thread 20:41 Krock the issue with it was that mapblocks have no version number, right? 20:42 nerzhul Krock, it's not a problem 20:42 nerzhul add the field and consider previous = v0 20:42 nerzhul or v1 20:42 sfan5 Krock, they do 20:42 nerzhul i did it on my server, i have 3 mapblocks versions 20:42 sofar zlib with avx acceleration by my team is what I want 20:43 nerzhul sofar, don't forget we are generic 20:43 sofar I don't want generic, I want to be able to keep using zlib (my OS version of it) 20:43 nerzhul we are not intel CPU only :p 20:43 sfan5 sofar, didn't some intel guy already PR the zlib repo with a sped-up impl? 20:44 sofar yes, not sure whether upstream has fully merged it or even cut a release, tho 20:44 nerzhul both brotli, zstd and lz4 are in repositories 20:44 sfan5 i don't think there's any usable version of either br or zstd in debian stable 20:45 sfan5 it's not a problem anyway as we can include the src 20:45 nerzhul https://packages.debian.org/stretch/brotli 20:45 nerzhul https://packages.debian.org/stretch/zstd 20:45 nerzhul i think it's better to have it in lib/ and permit to use system, it permits more recent compilers to optimize more 20:46 sfan5 hmm 20:46 sfan5 nerzhul, you're right zstd has one 20:46 sfan5 but libbrotli does not exist in anything but experimental 20:46 sfan5 not like this makes any difference for choosing a new algo 20:48 nore nerzhul: where should I put the tests do you think? 20:53 numzero Krock: Am I right you’re @SmallJoker? 20:54 Krock without the @ but yes 21:03 numzero “Crack grid” should be fixed now 21:12 paramat mgv6 still has floaty stuff occasionally due to cave carving 21:14 paramat thePalindrome has more floating bits :D 21:17 paramat but yes the basic terrain of mgv6 (without caves) does not create floaty stuff, due to being 2D noise 21:22 paramat i'm going to revert the 'make biomes, decorations, ores relative to water level' thing as i'm realising the problems it will cause, for a start 'get biome at pos' will be near impossible 21:22 paramat the first time i suggested it hmmmm refused it, now i realise he was right 21:27 paramat it's not much needed too 21:55 paramat commit 7000 approaches 22:01 Krock tenplus1 did commit 6666 22:01 paramat 4 dimensional beast 22:02 Krock indeed. pure evil 23:40 paramat #6425 rubenwardy please could you approve this revert? i would like to merge it in 2-3 hours after testing 23:40 ShadowBot https://github.com/minetest/minetest/issues/6425 -- Biome API: Revert biomes, decos, ores being relative to water level by paramat 23:41 rubenwardy is it a straight revert? 23:45 paramat yes precise revert 23:46 paramat but can't be an automated revert 23:48 paramat well, strictly there's one logic statement that i have simplified but with identical behaviour 23:54 paramat and it's better to have dedicated biomes for floatlands, that's intended