Time Nick Message 00:37 hmmmm well i'm not even sure what you're trying to accomplish there 00:38 Exio hmmmm: set the liquid_loop_max to 0 or 1 or something "insanely slow", generate a new map and you'll see a lot of "air holes" where the dungeons and caves get generated 00:40 hmmmm that's not supposed to happen for caves at least 00:41 Exio for the dungeons, it is a "one liner fix" (more like and if that isn't done in the air-generation but is in the wall-gen) 00:42 Exio for the caves, you are actually doing the if outside the if (large_cave) {} else {} for cave-v7 00:42 Exio but in the cave-v6 it is done inside the "small_cave" section 00:42 hmmmm for the dungeons, it's intentionally that way 00:43 Exio ah 00:43 hmmmm for caves v7, i did that for a different reason; i attempted to make small caves link together but it doesn't work too well 00:43 hmmmm instead what it does is eats through oceans and leaves 'bright spots' 00:44 Exio it looks like a glitch, that is why i though that wasn't intentional 00:44 celeron55 Exio: how much have you tested the patch? 00:45 hmmmm so, for v6 small caves at least, it can't replace content_ignore or else that'll happen, it can't replace water for the same exact reason (for a spot that's already generated), lava doesn't really matter too much, nothing is "buggy" when we're talking caves at the lava depth 00:45 hmmmm and not replacing air when it already places the air, that's so it doesn't mark places above ground as part of the caves 00:45 Exio celeron55: not a really lot but generated a lot of terrain (mainly was generating and checking if there was any "bubble") 00:45 hmmmm but since those flags are never used, it doesn't actually matter a whole lot 00:46 hmmmm as for large caves, water is placed where water would normally be generated so that check isn't needed there 00:46 celeron55 i mean, i shortly thought through the thing and it seems it makes sense for the water/lava part 00:46 celeron55 or at least water 00:46 hmmmm lava again doesn't matter too much, air doesn't matter either because it's placing air 00:47 Exio i just re-used the code that is behind 00:47 Exio i only would have checked for water :P 00:47 hmmmm yeah, bad idea. 00:48 Exio i thought the if was "correct" 00:48 hmmmm for dungeons you never check any of that, since you'd lose any neat effect of it looking destroyed by lava for example 00:49 celeron55 actually i think the ignore check is necessary too 00:49 hmmmm for what, dungeons? 00:49 celeron55 i am talking about this and only this https://github.com/EXio4/minetest/commit/08cd5809f4e8bd6ef253fbde0e5122440d507812 00:50 celeron55 dungeons are supposed to get generated on top of ignore, caves are not 00:50 hmmmm large caves need to be generated on top of ignore or else it'll get walled in by the next generated chunk and it'd look really odd 00:51 celeron55 is that something you have changed? 00:51 hmmmm no, that's the way it's been since forever 00:51 celeron55 umm... well i guess so then 00:51 hmmmm ... 00:52 hmmmm you're generating a large cave and it bumps up against the boundary, which happens to *not* be generated at this point 00:52 hmmmm since it's ignore, it wouldn't be written at all 00:52 hmmmm then that next chunk gets generated and places all stone nodes 00:52 hmmmm since it's still content ignore and not air 00:52 celeron55 i was thinking that maybe it previously generated all of the loaded ground, but that doesn't make that much sense 00:52 hmmmm so then when the player explores that cave, he'll notice a straight, smooth vertical wall cutting off the cave at the edge of a chunk boundary 00:53 hmmmm this is *undesirable* 00:53 celeron55 you don't need to explain that to me, it's not like i don't know 00:53 hmmmm it didn't seem that way with the "ummm.. well i guess so then" 00:55 celeron55 anyway, i think the large cave could check for water and lava and not replace them with air 00:55 celeron55 changing anything else than adding those two checks is probably a bad idea 00:58 Exio celeron55: https://github.com/EXio4/minetest/commit/b4c881eed0e0c17c64646fa9e93d55bb06fef566 ? 00:59 celeron55 ... 00:59 Exio wut? 00:59 hmmmm checking for water in large caves won't do anything, since water is placed in the same spot anyway 00:59 celeron55 you added two checks and *removed* two checks 00:59 celeron55 hmmmm: it does not? 01:00 hmmmm it does though 01:00 Exio well, whatever 01:00 celeron55 if that is the case, then there's no point in anything 01:00 Exio ^ exactly 01:00 hmmmm if (flooded && full_ymin < water_level && full_ymax > water_level) { 01:00 hmmmm vm->m_data[i] = (p.Y <= water_level) ? waternode : airnode; 01:00 celeron55 erm 01:00 celeron55 so does it or not? 01:01 hmmmm water occurs anywhere below water_level naturally 01:01 hmmmm the only place where this would make a difference is if flooded == false 01:01 hmmmm and that's with proller's version 01:01 celeron55 hmm ah, yes 01:02 celeron55 so the thing Exio is trying to fix is actually taken care of already 01:02 Exio then i guess i have a commit somewhere hidden in my mt 01:02 hmmmm well what he was trying to fix is a visible problem 01:02 hmmmm you can notice it apparently by disabling liquid transforming 01:03 hmmmm well; 'problem' here is being used leniently here 01:03 celeron55 the air bubbles exio is seeing could be large caves that were generated on top of ignore, and the water not getting generated on that in the neighboring chunk 01:04 celeron55 or the bottom parts of small caves at the ocean floor (which will be air) 01:04 celeron55 probably both 01:04 hmmmm nevertheless; 'problems' with map generation are pretty subjective and don't affect much 01:04 celeron55 both are hard to fix and it's probably just best to let the water flow there with the existing flowing mechanisms... 01:05 Exio i guess 01:05 Exio then proller will rage at you for his ocean flowing away ;P 01:05 hmmmm i thought it replenishes itself if the water around it has a sufficient mass to prevent that sort of thing 01:06 Exio hmmmm: with finite_liquid(s) 01:06 hmmmm yeah 01:06 hmmmm the so-called 'relax' setting or something like that; not really familiar with how finite liquids work 01:07 Exio well, i'm not really sure 01:08 celeron55 once again we could make a "perfect" world by always generating water and never air below ocean level, but that would be boring as hell 01:08 celeron55 it's always a compromise 01:20 hmmmm making it better would involve changing some water physics... i noticed from watching minecraft videos that if a large space of air is surrounded by a big enough amount of water, it'll flood fill it all immediately 01:21 hmmmm if we did something like that there'd be less bubbles 01:21 Exio something like making "water flow" on mapgen time? 01:21 hmmmm no... nothing like that 16:45 proller bug in collision 8( 16:47 proller sand stuck because collision with bottom-left(maybe right top bottom) block, with empty bottom 16:47 proller and falling check empty bottom and skip it 16:48 proller can reproduce on float islands with falling sand 16:55 celeron55 well that is 100% likely a floating point precision problem and the only way to fix it will be making the collision box smaller 16:55 celeron55 or, well... probably some other way too, which is a variation of the same fix 17:04 proller im trying debug and find how smaller 18:00 Taoki http://forum.minetest.net/viewtopic.php?pid=97450#p97450 18:13 Taoki http://forum.minetest.net/viewtopic.php?pid=97455#p97455 Better screenshots 18:14 VanessaE Taoki: beautiful. 18:14 Taoki Thanks 18:14 Taoki http://i42.tinypic.com/2hzsvgn.png ---- http://i44.tinypic.com/jkevkg.png 18:28 proller falling collision fix - https://github.com/proller/minetest/commit/9c2026799dec58aab4a465e8aa80613f41bdf4e9 any objections? 18:36 hmmmm isn't 0.03125 way too big for a floating point error? 18:36 hmmmm should be more like 0.00004 or something 18:39 proller its not very floating, its near algo error 18:41 proller bunches of sand fall better than before 18:41 proller more smooth 18:41 proller without micro jumps 18:42 proller will commit! 18:43 proller if it broke something - easy to back via #define COLL_ZERO 0 20:17 Taoki PilzAdam: http://forum.minetest.net/viewtopic.php?pid=97481#p97481 Can you please make another build? I just amended the commit 60 seconds ago to fix the yellow fog 20:17 PilzAdam hm? 20:17 Taoki Well not really fix... made it a little bit more orange tinted. But overall it gives a nicer feel, so it's a little better 20:18 Taoki erm... 20:18 Taoki PilzAdam: Sorry, that's sfan5. He has the same avatar as you on the Forum :P 20:18 PilzAdam no, he has not 20:18 PilzAdam he has an apple in the tree IIRC 20:18 Taoki Ah, it's the default avatar then 20:18 Taoki strange, I see the default one 20:18 PilzAdam there is no default avatar 20:18 sfan5 Taoki: ok 20:20 Taoki But yeah, I got it looking as good as it can look at least for the time being 20:20 sfan5 it looks awesome 20:20 Taoki Thanks 20:21 PilzAdam proller, nice! you broke unittests! 20:21 sfan5 hah 20:21 Taoki PilzAdam: If you wish, see latest screenshots and try the code. I'd like to know if you agree. Note that it's unlikely I can improve anything any more howevber 20:21 PilzAdam proller, revert that commit 20:21 Taoki What's unittests? 20:21 proller hm 20:21 proller wait.. 20:22 sfan5 Taoki: some tests that ensure e.g. that the configuration saving of strings/numbers/floats/booleans etc. works 20:23 Taoki ok :) 20:23 Taoki Taking opinions on the fog color for 1 - 2 more hours or so, then I'll pull request it (not like I expect that to make a difference ;] ) 20:23 Taoki **fog code 20:26 PilzAdam Taoki, clouds and fog have different colors 20:26 Taoki They can't always be matched 20:26 PilzAdam they should 20:26 PilzAdam otherwise it looks weird 20:26 PilzAdam e.g. the clouds are orange and the fog is yellow 20:27 Taoki I didn't even notice. Nor have any way to do that I can think of 20:27 Taoki Fig is now orange-ish too 20:27 Taoki **fog 20:28 Taoki Also, after the sun has fully set, there's a discrete purple-red tint at the horrizon. No way I can match the fog to that. And it's the most beautiful part of the new colors 20:29 Taoki erm, match the clouds to that x.x 20:30 PilzAdam your patch doesnt look good as it is currently 20:30 proller tests must be fixed 20:30 Taoki What's not good about it? 20:30 PilzAdam the colors dont match the ambient 20:30 PilzAdam maybe try using the cloud color 20:31 Taoki Cloud colors would ruin it. And sunrise / sunset is meant to not match the sky color 20:31 PilzAdam proller, also I dont see that any other core dev has agreed to your commit 20:32 Taoki The fog colors tend to match cloud colors to a small extent though... at least the orange / yellow part 20:32 proller PilzAdam, nobody against 20:32 Taoki Also, fog is added together with the horrizon. The horrizon didn't match cloud colors before either 20:33 PilzAdam proller, the rule is that somebody has to agree on this 20:33 PilzAdam also hmmmm was against it 20:33 Taoki Nor does this feature do so in Minecraft. Not that we need to copy how it's done there, but no one complains about it there either 20:33 Taoki (I think in MC clouds are always white) 20:34 proller PilzAdam, its was remark 20:34 Taoki Anyway, those aren't things I can change. I'll pull request it and see what the other devs think 20:35 Taoki Not surprised if this also goes to the trash can like many contributions though, for all sorts of reasons. When after 3 days of working on it I managed to add an awesome featoure to Minetest (hardly even as it is now) 21:07 hmmmm ? 21:07 hmmmm the latest commit? i wasn't against it, but we can't have things breaking unit tests